Executive summary
crmne/ruby_llm falls short of our standard for production, in part because it has pipeline tools referenced by a changeable name. It scores 62 out of 100 and is rated Adequate, and 53 of its 54 critical findings are security problems in the build pipeline and in its code.
Security and Compliance, at 55, is the weakest lens and counts most in the score. The best first step for the work it takes is stopping the eight pipeline steps that hand secrets to unpinned tools, worth about 3 points. Replacing the weak hashes comes next, worth about 5 points. Raising Security and Compliance to 70 would lift the score to about 67.
Thirty-three pipeline tools are referenced by a changeable name, so the build pipeline can be tampered with. Thirteen secrets can leak from the build pipeline, and five places in the code run text as code, which an attacker could use. Pinning each tool to an exact version, limiting each secret to its own step and replacing text run as code with direct calls would close these risks. Passing outside input through environment variables would fix the pipeline step that puts outside input straight into a shell command, and fixing the weaknesses the scanner points to would deal with the security weakness in the code.
The product code is 49,973 lines, about €130,000 to rebuild.
If Security and Compliance rose from 55 to 70, the lowest score rated Strong, the score would rise from 62 to about 67.
New since the last survey: 32 findings
- Cyclomatic Complexity · Chat.render_payload (cyclomatic 17) lib/ruby_llm/protocols/gemini/chat.rb
- God Classes · TooManyMethods: CodeExecution lib/ruby_llm/code_execution.rb
- God Classes · TooManyMethods: Tool lib/ruby_llm/tool.rb
- God Classes · TooManyMethods: Message lib/ruby_llm/message.rb
- God Classes · FileTooLong: ruby_llm/provider.rb lib/ruby_llm/provider.rb
- Code Duplication · Duplicated block (13–14 lines × 2) lib/ruby_llm/image.rb
- Code Duplication · Duplicated block (9 lines × 2) lib/ruby_llm/protocols/chat_completions/streaming.rb
- Code Duplication · Duplicated block (24 lines × 2) lib/ruby_llm/providers/hetzner/chat.rb
- Code Duplication · Duplicated block (9 lines × 3) lib/ruby_llm/file_store/tools/search.rb
- Code Duplication · Duplicated block (5 lines × 2) lib/ruby_llm/prompt.rb
- Code Duplication · Duplicated block (5 lines × 2) lib/ruby_llm/protocols/cohere/tools.rb
- Code Duplication · Duplicated block (8 lines × 2) lib/ruby_llm/protocols/converse/chat.rb
- Cohesion (LCOM4) · Low cohesion: Protocol (LCOM4 7) lib/ruby_llm/protocol.rb
- Cohesion (LCOM4) · Low cohesion: Task (LCOM4 4) lib/ruby_llm/mcp/task.rb
- Test Reliability · Flaky test: RubyLLM::Chat provider mapping lets provider options override the mapped value
- Churn × Complexity Hotspots · Hotspot: lib/ruby_llm/protocols/gemini/chat.rb lib/ruby_llm/protocols/gemini/chat.rb
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Static Analysis (SAST) · REDACTED
- Visual & motion safety · Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css
- Visual & motion safety · Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css
- Visual & motion safety · Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css
- Visual & motion safety · Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css
- Visual & motion safety · Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css
Fixed since the last survey: 9 findings
- Code Duplication · Duplicated block (11–12 lines × 3) lib/ruby_llm/image.rb
- Code Duplication · Duplicated block (10–11 lines × 2) lib/ruby_llm/protocols/chat_completions/streaming.rb
- Code Duplication · Duplicated block (8 lines × 3) lib/ruby_llm/protocols/anthropic/chat.rb
- Code Duplication · Duplicated block (5 lines × 2) lib/ruby_llm/protocols/cohere/tools.rb
- Code Duplication · Duplicated block (5 lines × 2) lib/ruby_llm/protocols/chat_completions/chat.rb
- Cohesion (LCOM4) · Low cohesion: Protocol (LCOM4 6) lib/ruby_llm/protocol.rb
- Static Analysis (SAST) · High: REDACTED REDACTED
- Static Analysis (SAST) · High: REDACTED REDACTED
- Change Coupling · Change coupling: agent.rb ↔ REDACTED lib/ruby_llm/agent.rb
Rebuild cost: about €130,000 · Estimate
| Cost to rebuild | €44,000 to €220,000, for 739 to 2,345 hours of work, about 0.9 person-years for 1 to 2 engineers |
|---|---|
| Difficulty of the problem | Standard, judged from the kind of software and how it is built. Code that solves a harder problem, such as payments or medical bookings, costs more per line to write. |
| Quality factor | ×0.8 at a score of 62. Better code takes more work to write, so the factor is ×1.0 at a score of 70, lower below it and higher above it. |
| Size | Medium: 49,973 lines of product code |
| What it is not | Not a quotation: nobody has offered to build it for that. Not a valuation: it says nothing about what the business is worth. Not a certified figure: no professional has signed it as one. |
Top priorities
Diagnosis: what the measurements show together
Architecture: which parts of the code depend on which
| uses → | 1 Batch | 2 Chat | 3 CodeExecution | 4 Providers.Bedrock | 5 Support | 6 Accounting.Usage | 7 MCP | 8 Protocol | 9 Protocols.Anthropic | 10 Protocols.Cohere | 11 Transport | 12 Evaluation | 13 RubyLLM | 14 ActiveRecord | 15 Protocols.Bedrock | 16 Protocols.ElevenLabs | 17 Protocols.Gemini | 18 Protocols.OpenAI | 19 Providers | 20 Protocols.ChatCompletions | 21 Protocols.Converse | 22 Protocols.Interactions | 23 Protocols.Responses | 24 Protocols | 25 Protocols.GPUStack | 26 Protocols.InvokeModel | 27 Protocols.Mistral | 28 Protocols.OpenRouter | 29 Protocols.VertexAI | 30 Protocols.XAI | 31 Providers.Azure | 32 Providers.Bedrock.Mantle | 33 Providers.Meta | 34 Providers.Ollama | 35 Providers.OpenAI | 36 Providers.GPUStack | 37 Providers.Mistral | 38 Providers.OpenRouter | 39 Providers.XAI | 40 Providers.VertexAI |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 Batch | ||||||||||||||||||||||||||||||||||||||||
| 2 Chat | ||||||||||||||||||||||||||||||||||||||||
| 3 CodeExecution | ||||||||||||||||||||||||||||||||||||||||
| 4 Providers.Bedrock | ||||||||||||||||||||||||||||||||||||||||
| 5 Support | ||||||||||||||||||||||||||||||||||||||||
| 6 Accounting.Usage | 1 | |||||||||||||||||||||||||||||||||||||||
| 7 MCP | 5 | 5 | ||||||||||||||||||||||||||||||||||||||
| 8 Protocol | 1 | |||||||||||||||||||||||||||||||||||||||
| 9 Protocols.Anthropic | 1 | 1 | ||||||||||||||||||||||||||||||||||||||
| 10 Protocols.Cohere | 1 | 1 | ||||||||||||||||||||||||||||||||||||||
| 11 Transport | 1 | 1 | ||||||||||||||||||||||||||||||||||||||
| 12 Evaluation | 4 | 2 | 1 | |||||||||||||||||||||||||||||||||||||
| 13 RubyLLM | 2 | 1 | 43 | 10 | 1 | 5 | 1 | 2 | ||||||||||||||||||||||||||||||||
| 14 ActiveRecord | 1 | |||||||||||||||||||||||||||||||||||||||
| 15 Protocols.Bedrock | 1 | 3 | 1 | |||||||||||||||||||||||||||||||||||||
| 16 Protocols.ElevenLabs | 1 | 1 | ||||||||||||||||||||||||||||||||||||||
| 17 Protocols.Gemini | 1 | 1 | 1 | |||||||||||||||||||||||||||||||||||||
| 18 Protocols.OpenAI | 1 | 1 | 1 | |||||||||||||||||||||||||||||||||||||
| 19 Providers | 2 | 23 | ||||||||||||||||||||||||||||||||||||||
| 20 Protocols.ChatCompletions | 2 | |||||||||||||||||||||||||||||||||||||||
| 21 Protocols.Converse | 1 | |||||||||||||||||||||||||||||||||||||||
| 22 Protocols.Interactions | 1 | |||||||||||||||||||||||||||||||||||||||
| 23 Protocols.Responses | 1 | |||||||||||||||||||||||||||||||||||||||
| 24 Protocols | 7 | 11 | 11 | 4 | 13 | 12 | 4 | 6 | 6 | |||||||||||||||||||||||||||||||
| 25 Protocols.GPUStack | 1 | |||||||||||||||||||||||||||||||||||||||
| 26 Protocols.InvokeModel | 1 | 5 | ||||||||||||||||||||||||||||||||||||||
| 27 Protocols.Mistral | 2 | |||||||||||||||||||||||||||||||||||||||
| 28 Protocols.OpenRouter | 1 | 2 | ||||||||||||||||||||||||||||||||||||||
| 29 Protocols.VertexAI | 1 | 2 | 2 | |||||||||||||||||||||||||||||||||||||
| 30 Protocols.XAI | 1 | |||||||||||||||||||||||||||||||||||||||
| 31 Providers.Azure | 1 | 3 | ||||||||||||||||||||||||||||||||||||||
| 32 Providers.Bedrock.Mantle | 3 | 3 | ||||||||||||||||||||||||||||||||||||||
| 33 Providers.Meta | 1 | 1 | ||||||||||||||||||||||||||||||||||||||
| 34 Providers.Ollama | 2 | |||||||||||||||||||||||||||||||||||||||
| 35 Providers.OpenAI | 2 | 1 | ||||||||||||||||||||||||||||||||||||||
| 36 Providers.GPUStack | 1 | 1 | 2 | |||||||||||||||||||||||||||||||||||||
| 37 Providers.Mistral | 1 | 2 | ||||||||||||||||||||||||||||||||||||||
| 38 Providers.OpenRouter | 1 | 2 | 1 | |||||||||||||||||||||||||||||||||||||
| 39 Providers.XAI | 1 | 2 | 4 | |||||||||||||||||||||||||||||||||||||
| 40 Providers.VertexAI | 2 | 5 | 4 |
6→5 Accounting.Usage uses Support✕
- Connections
- 1 pair of types, where a type in Accounting.Usage uses a type in Support.
- The types involved
RubyLLM.Accounting.Usage.Entry→RubyLLM.Support.Inspectable
- Direction
- Accounting.Usage uses Support, so a change to Support can break Accounting.Usage, while a change to Accounting.Usage can't break Support.
- Position
- Below the diagonal: Accounting.Usage uses a module listed above it, which is the healthy direction.
7→5 MCP uses Support✕
- Connections
- 5 pairs of types, where a type in MCP uses a type in Support.
- The types involved
RubyLLM.MCP.Prompt→RubyLLM.Support.InspectableRubyLLM.MCP.Resource→RubyLLM.Support.InspectableRubyLLM.MCP.ResourceTemplate→RubyLLM.Support.InspectableRubyLLM.MCP.Result→RubyLLM.Support.InspectableRubyLLM.MCP.Tool→RubyLLM.Support.Inspectable
- Direction
- MCP uses Support, so a change to Support can break MCP, while a change to MCP can't break Support.
- Position
- Below the diagonal: MCP uses a module listed above it, which is the healthy direction.
7→13 MCP uses RubyLLM part of a loop✕
- Connections
- 5 pairs of types, where a type in MCP uses a type in RubyLLM.
- The types involved
RubyLLM.MCP.Error→RubyLLM.ErrorRubyLLM.MCP.InputRequiredError→RubyLLM.InputRequiredErrorRubyLLM.MCP.InsufficientScopeError→RubyLLM.ForbiddenErrorRubyLLM.MCP.Task→RubyLLM.TaskRubyLLM.MCP.Tool→RubyLLM.Tool
- Direction
- MCP uses RubyLLM, so a change to RubyLLM can break MCP, while a change to MCP can't break RubyLLM.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
8→13 Protocol uses RubyLLM part of a loop✕
- Connections
- 1 pair of types, where a type in Protocol uses a type in RubyLLM.
- The types involved
RubyLLM.Protocol.StreamAccumulator→RubyLLM.Tokens
- Direction
- Protocol uses RubyLLM, so a change to RubyLLM can break Protocol, while a change to Protocol can't break RubyLLM.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
9→1 Protocols.Anthropic uses Batch✕
- Connections
- 1 pair of types, where a type in Protocols.Anthropic uses a type in Batch.
- The types involved
RubyLLM.Protocols.Anthropic.Batches→RubyLLM.Batch.Helpers
- Direction
- Protocols.Anthropic uses Batch, so a change to Batch can break Protocols.Anthropic, while a change to Protocols.Anthropic can't break Batch.
- Position
- Below the diagonal: Protocols.Anthropic uses a module listed above it, which is the healthy direction.
9→24 Protocols.Anthropic uses Protocols part of a loop✕
- Connections
- 1 pair of types, where a type in Protocols.Anthropic uses a type in Protocols.
- The types involved
RubyLLM.Protocols.Anthropic.Files→RubyLLM.Protocols.Files
- Direction
- Protocols.Anthropic uses Protocols, so a change to Protocols can break Protocols.Anthropic, while a change to Protocols.Anthropic can't break Protocols.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
10→1 Protocols.Cohere uses Batch✕
- Connections
- 1 pair of types, where a type in Protocols.Cohere uses a type in Batch.
- The types involved
RubyLLM.Protocols.Cohere.Batches→RubyLLM.Batch.Helpers
- Direction
- Protocols.Cohere uses Batch, so a change to Batch can break Protocols.Cohere, while a change to Protocols.Cohere can't break Batch.
- Position
- Below the diagonal: Protocols.Cohere uses a module listed above it, which is the healthy direction.
10→24 Protocols.Cohere uses Protocols part of a loop✕
- Connections
- 1 pair of types, where a type in Protocols.Cohere uses a type in Protocols.
- The types involved
RubyLLM.Protocols.Cohere.Datasets→RubyLLM.Protocols.Files
- Direction
- Protocols.Cohere uses Protocols, so a change to Protocols can break Protocols.Cohere, while a change to Protocols.Cohere can't break Protocols.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
11→5 Transport uses Support✕
- Connections
- 1 pair of types, where a type in Transport uses a type in Support.
- The types involved
RubyLLM.Transport.Connection→RubyLLM.Support.Inspectable
- Direction
- Transport uses Support, so a change to Support can break Transport, while a change to Transport can't break Support.
- Position
- Below the diagonal: Transport uses a module listed above it, which is the healthy direction.
11→13 Transport uses RubyLLM part of a loop✕
- Connections
- 1 pair of types, where a type in Transport uses a type in RubyLLM.
- The types involved
RubyLLM.Transport.WebsocketConnection→RubyLLM.Error
- Direction
- Transport uses RubyLLM, so a change to RubyLLM can break Transport, while a change to Transport can't break RubyLLM.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
12→5 Evaluation uses Support✕
- Connections
- 4 pairs of types, where a type in Evaluation uses a type in Support.
- The types involved
RubyLLM.Evaluation.Case→RubyLLM.Support.InspectableRubyLLM.Evaluation.Report→RubyLLM.Support.InspectableRubyLLM.Evaluation.Result→RubyLLM.Support.InspectableRubyLLM.Evaluation.Trial→RubyLLM.Support.Inspectable
- Direction
- Evaluation uses Support, so a change to Support can break Evaluation, while a change to Evaluation can't break Support.
- Position
- Below the diagonal: Evaluation uses a module listed above it, which is the healthy direction.
12→6 Evaluation uses Accounting.Usage✕
- Connections
- 2 pairs of types, where a type in Evaluation uses a type in Accounting.Usage.
- The types involved
RubyLLM.Evaluation.Report→RubyLLM.Accounting.Usage.ResultRubyLLM.Evaluation.Trial→RubyLLM.Accounting.Usage.Result
- Direction
- Evaluation uses Accounting.Usage, so a change to Accounting.Usage can break Evaluation, while a change to Evaluation can't break Accounting.Usage.
- Position
- Below the diagonal: Evaluation uses a module listed above it, which is the healthy direction.
12→13 Evaluation uses RubyLLM part of a loop✕
- Connections
- 1 pair of types, where a type in Evaluation uses a type in RubyLLM.
- The types involved
RubyLLM.Evaluation.Reviewer→RubyLLM.Agent
- Direction
- Evaluation uses RubyLLM, so a change to RubyLLM can break Evaluation, while a change to Evaluation can't break RubyLLM.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
13→2 RubyLLM uses Chat✕
- Connections
- 2 pairs of types, where a type in RubyLLM uses a type in Chat.
- The types involved
RubyLLM.Chat→RubyLLM.Chat.ToolSearchRubyLLM.Chat→RubyLLM.Chat.ToolSequences
- Direction
- RubyLLM uses Chat, so a change to Chat can break RubyLLM, while a change to RubyLLM can't break Chat.
- Position
- Below the diagonal: RubyLLM uses a module listed above it, which is the healthy direction.
13→3 RubyLLM uses CodeExecution✕
- Connections
- 1 pair of types, where a type in RubyLLM uses a type in CodeExecution.
- The types involved
RubyLLM.CodeExecution→RubyLLM.CodeExecution.Packages
- Direction
- RubyLLM uses CodeExecution, so a change to CodeExecution can break RubyLLM, while a change to RubyLLM can't break CodeExecution.
- Position
- Below the diagonal: RubyLLM uses a module listed above it, which is the healthy direction.
13→5 RubyLLM uses Support✕
- Connections
- 43 pairs of types, where a type in RubyLLM uses a type in Support. Showing 25 of them; the rest are in namespace-graph.json in this report's bundle.
- The types involved
RubyLLM.Attachment→RubyLLM.Support.InspectableRubyLLM.Batch→RubyLLM.Support.InspectableRubyLLM.CachedContent→RubyLLM.Support.InspectableRubyLLM.Chat→RubyLLM.Support.InspectableRubyLLM.Choice→RubyLLM.Support.InspectableRubyLLM.Citation→RubyLLM.Support.InspectableRubyLLM.ContentFilter→RubyLLM.Support.InspectableRubyLLM.Cost→RubyLLM.Support.InspectableRubyLLM.DownloadedFile→RubyLLM.Support.InspectableRubyLLM.Embedding→RubyLLM.Support.InspectableRubyLLM.EmbeddingRequest→RubyLLM.Support.InspectableRubyLLM.Fallback→RubyLLM.Support.InspectableRubyLLM.Image→RubyLLM.Support.InspectableRubyLLM.InputRequest→RubyLLM.Support.InspectableRubyLLM.Judgment→RubyLLM.Support.InspectableRubyLLM.MCP→RubyLLM.Support.InspectableRubyLLM.Message→RubyLLM.Support.InspectableRubyLLM.Model→RubyLLM.Support.InspectableRubyLLM.Moderation→RubyLLM.Support.InspectableRubyLLM.OCR→RubyLLM.Support.InspectableRubyLLM.Probability→RubyLLM.Support.InspectableRubyLLM.Progress→RubyLLM.Support.InspectableRubyLLM.Provider→RubyLLM.Support.InspectableRubyLLM.QueuedMessage→RubyLLM.Support.InspectableRubyLLM.RequestShape→RubyLLM.Support.Inspectable
- Direction
- RubyLLM uses Support, so a change to Support can break RubyLLM, while a change to RubyLLM can't break Support.
- Position
- Below the diagonal: RubyLLM uses a module listed above it, which is the healthy direction.
13→6 RubyLLM uses Accounting.Usage✕
- Connections
- 10 pairs of types, where a type in RubyLLM uses a type in Accounting.Usage.
- The types involved
RubyLLM.Embedding→RubyLLM.Accounting.Usage.ResultRubyLLM.Image→RubyLLM.Accounting.Usage.ResultRubyLLM.Judgment→RubyLLM.Accounting.Usage.ResultRubyLLM.Message→RubyLLM.Accounting.Usage.ResultRubyLLM.Moderation→RubyLLM.Accounting.Usage.ResultRubyLLM.OCR→RubyLLM.Accounting.Usage.ResultRubyLLM.Protocol→RubyLLM.Accounting.Usage.TrackerRubyLLM.Rerank→RubyLLM.Accounting.Usage.ResultRubyLLM.Speech→RubyLLM.Accounting.Usage.ResultRubyLLM.Transcription→RubyLLM.Accounting.Usage.Result
- Direction
- RubyLLM uses Accounting.Usage, so a change to Accounting.Usage can break RubyLLM, while a change to RubyLLM can't break Accounting.Usage.
- Position
- Below the diagonal: RubyLLM uses a module listed above it, which is the healthy direction.
13→7 RubyLLM uses MCP✕
- Connections
- 1 pair of types, where a type in RubyLLM uses a type in MCP.
- The types involved
RubyLLM.Chat→RubyLLM.MCP.Collection
- Direction
- RubyLLM uses MCP, so a change to MCP can break RubyLLM, while a change to RubyLLM can't break MCP.
- Position
- Below the diagonal: RubyLLM uses a module listed above it, which is the healthy direction.
13→8 RubyLLM uses Protocol✕
- Connections
- 5 pairs of types, where a type in RubyLLM uses a type in Protocol.
- The types involved
RubyLLM.Protocol→RubyLLM.Protocol.BinaryStreamingRubyLLM.Protocol→RubyLLM.Protocol.NativeContentRubyLLM.Protocol→RubyLLM.Protocol.RequestShapesRubyLLM.Protocol→RubyLLM.Protocol.StreamingRubyLLM.Protocol→RubyLLM.Protocol.ToolNamespaces
- Direction
- RubyLLM uses Protocol, so a change to Protocol can break RubyLLM, while a change to RubyLLM can't break Protocol.
- Position
- Below the diagonal: RubyLLM uses a module listed above it, which is the healthy direction.
13→11 RubyLLM uses Transport✕
- Connections
- 1 pair of types, where a type in RubyLLM uses a type in Transport.
- The types involved
RubyLLM.Provider→RubyLLM.Transport.Connection
- Direction
- RubyLLM uses Transport, so a change to Transport can break RubyLLM, while a change to RubyLLM can't break Transport.
- Position
- Below the diagonal: RubyLLM uses a module listed above it, which is the healthy direction.
13→12 RubyLLM uses Evaluation✕
- Connections
- 2 pairs of types, where a type in RubyLLM uses a type in Evaluation.
- The types involved
RubyLLM.Evaluation→RubyLLM.Evaluation.EvaluatorRubyLLM.Evaluation→RubyLLM.Evaluation.Evidence
- Direction
- RubyLLM uses Evaluation, so a change to Evaluation can break RubyLLM, while a change to RubyLLM can't break Evaluation.
- Position
- Below the diagonal: RubyLLM uses a module listed above it, which is the healthy direction.
14→13 ActiveRecord uses RubyLLM✕
- Connections
- 1 pair of types, where a type in ActiveRecord uses a type in RubyLLM.
- The types involved
RubyLLM.ActiveRecord.FileStore→RubyLLM.FileStore
- Direction
- ActiveRecord uses RubyLLM, so a change to RubyLLM can break ActiveRecord, while a change to ActiveRecord can't break RubyLLM.
- Position
- Below the diagonal: ActiveRecord uses a module listed above it, which is the healthy direction.
15→1 Protocols.Bedrock uses Batch✕
- Connections
- 1 pair of types, where a type in Protocols.Bedrock uses a type in Batch.
- The types involved
RubyLLM.Protocols.Bedrock.Batches→RubyLLM.Batch.Helpers
- Direction
- Protocols.Bedrock uses Batch, so a change to Batch can break Protocols.Bedrock, while a change to Protocols.Bedrock can't break Batch.
- Position
- Below the diagonal: Protocols.Bedrock uses a module listed above it, which is the healthy direction.
15→13 Protocols.Bedrock uses RubyLLM✕
- Connections
- 3 pairs of types, where a type in Protocols.Bedrock uses a type in RubyLLM.
- The types involved
RubyLLM.Protocols.Bedrock.AsyncVideos→RubyLLM.ProtocolRubyLLM.Protocols.Bedrock.Guardrails→RubyLLM.ProtocolRubyLLM.Protocols.Bedrock.Rerank→RubyLLM.Protocol
- Direction
- Protocols.Bedrock uses RubyLLM, so a change to RubyLLM can break Protocols.Bedrock, while a change to Protocols.Bedrock can't break RubyLLM.
- Position
- Below the diagonal: Protocols.Bedrock uses a module listed above it, which is the healthy direction.
15→24 Protocols.Bedrock uses Protocols part of a loop✕
- Connections
- 1 pair of types, where a type in Protocols.Bedrock uses a type in Protocols.
- The types involved
RubyLLM.Protocols.Bedrock.Files→RubyLLM.Protocols.Files
- Direction
- Protocols.Bedrock uses Protocols, so a change to Protocols can break Protocols.Bedrock, while a change to Protocols.Bedrock can't break Protocols.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
16→13 Protocols.ElevenLabs uses RubyLLM✕
- Connections
- 1 pair of types, where a type in Protocols.ElevenLabs uses a type in RubyLLM.
- The types involved
RubyLLM.Protocols.ElevenLabs.Flows→RubyLLM.Protocol
- Direction
- Protocols.ElevenLabs uses RubyLLM, so a change to RubyLLM can break Protocols.ElevenLabs, while a change to Protocols.ElevenLabs can't break RubyLLM.
- Position
- Below the diagonal: Protocols.ElevenLabs uses a module listed above it, which is the healthy direction.
16→24 Protocols.ElevenLabs uses Protocols part of a loop✕
- Connections
- 1 pair of types, where a type in Protocols.ElevenLabs uses a type in Protocols.
- The types involved
RubyLLM.Protocols.ElevenLabs.Assets→RubyLLM.Protocols.Files
- Direction
- Protocols.ElevenLabs uses Protocols, so a change to Protocols can break Protocols.ElevenLabs, while a change to Protocols.ElevenLabs can't break Protocols.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
17→1 Protocols.Gemini uses Batch✕
- Connections
- 1 pair of types, where a type in Protocols.Gemini uses a type in Batch.
- The types involved
RubyLLM.Protocols.Gemini.Batches→RubyLLM.Batch.Helpers
- Direction
- Protocols.Gemini uses Batch, so a change to Batch can break Protocols.Gemini, while a change to Protocols.Gemini can't break Batch.
- Position
- Below the diagonal: Protocols.Gemini uses a module listed above it, which is the healthy direction.
17→13 Protocols.Gemini uses RubyLLM✕
- Connections
- 1 pair of types, where a type in Protocols.Gemini uses a type in RubyLLM.
- The types involved
RubyLLM.Protocols.Gemini.LiveTranscription→RubyLLM.Protocol
- Direction
- Protocols.Gemini uses RubyLLM, so a change to RubyLLM can break Protocols.Gemini, while a change to Protocols.Gemini can't break RubyLLM.
- Position
- Below the diagonal: Protocols.Gemini uses a module listed above it, which is the healthy direction.
17→24 Protocols.Gemini uses Protocols part of a loop✕
- Connections
- 1 pair of types, where a type in Protocols.Gemini uses a type in Protocols.
- The types involved
RubyLLM.Protocols.Gemini.Files→RubyLLM.Protocols.Files
- Direction
- Protocols.Gemini uses Protocols, so a change to Protocols can break Protocols.Gemini, while a change to Protocols.Gemini can't break Protocols.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
18→1 Protocols.OpenAI uses Batch✕
- Connections
- 1 pair of types, where a type in Protocols.OpenAI uses a type in Batch.
- The types involved
RubyLLM.Protocols.OpenAI.Batches→RubyLLM.Batch.Helpers
- Direction
- Protocols.OpenAI uses Batch, so a change to Batch can break Protocols.OpenAI, while a change to Protocols.OpenAI can't break Batch.
- Position
- Below the diagonal: Protocols.OpenAI uses a module listed above it, which is the healthy direction.
18→13 Protocols.OpenAI uses RubyLLM✕
- Connections
- 1 pair of types, where a type in Protocols.OpenAI uses a type in RubyLLM.
- The types involved
RubyLLM.Protocols.OpenAI.Decisions→RubyLLM.Protocol
- Direction
- Protocols.OpenAI uses RubyLLM, so a change to RubyLLM can break Protocols.OpenAI, while a change to Protocols.OpenAI can't break RubyLLM.
- Position
- Below the diagonal: Protocols.OpenAI uses a module listed above it, which is the healthy direction.
18→24 Protocols.OpenAI uses Protocols part of a loop✕
- Connections
- 1 pair of types, where a type in Protocols.OpenAI uses a type in Protocols.
- The types involved
RubyLLM.Protocols.OpenAI.Files→RubyLLM.Protocols.Files
- Direction
- Protocols.OpenAI uses Protocols, so a change to Protocols can break Protocols.OpenAI, while a change to Protocols.OpenAI can't break Protocols.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
19→4 Providers uses Providers.Bedrock✕
- Connections
- 2 pairs of types, where a type in Providers uses a type in Providers.Bedrock.
- The types involved
RubyLLM.Providers.Bedrock→RubyLLM.Providers.Bedrock.AuthRubyLLM.Providers.Bedrock→RubyLLM.Providers.Bedrock.Models
- Direction
- Providers uses Providers.Bedrock, so a change to Providers.Bedrock can break Providers, while a change to Providers can't break Providers.Bedrock.
- Position
- Below the diagonal: Providers uses a module listed above it, which is the healthy direction.
19→13 Providers uses RubyLLM✕
- Connections
- 23 pairs of types, where a type in Providers uses a type in RubyLLM.
- The types involved
RubyLLM.Providers.Anthropic→RubyLLM.ProviderRubyLLM.Providers.Azure→RubyLLM.ProviderRubyLLM.Providers.Bedrock→RubyLLM.ProviderRubyLLM.Providers.Cohere→RubyLLM.ProviderRubyLLM.Providers.DeepSeek→RubyLLM.ProviderRubyLLM.Providers.Deepgram→RubyLLM.ProviderRubyLLM.Providers.ElevenLabs→RubyLLM.ProviderRubyLLM.Providers.GPUStack→RubyLLM.ProviderRubyLLM.Providers.Gemini→RubyLLM.ProviderRubyLLM.Providers.Hetzner→RubyLLM.ProviderRubyLLM.Providers.Meta→RubyLLM.ProviderRubyLLM.Providers.MiniMax→RubyLLM.ProviderRubyLLM.Providers.Mistral→RubyLLM.ProviderRubyLLM.Providers.Moonshot→RubyLLM.ProviderRubyLLM.Providers.Ollama→RubyLLM.ProviderRubyLLM.Providers.OllamaCloud→RubyLLM.ProviderRubyLLM.Providers.OpenAI→RubyLLM.ProviderRubyLLM.Providers.OpenRouter→RubyLLM.ProviderRubyLLM.Providers.Perplexity→RubyLLM.ProviderRubyLLM.Providers.TypeSafe→RubyLLM.ProviderRubyLLM.Providers.VertexAI→RubyLLM.ProviderRubyLLM.Providers.XAI→RubyLLM.ProviderRubyLLM.Providers.ZAI→RubyLLM.Provider
- Direction
- Providers uses RubyLLM, so a change to RubyLLM can break Providers, while a change to Providers can't break RubyLLM.
- Position
- Below the diagonal: Providers uses a module listed above it, which is the healthy direction.
20→18 Protocols.ChatCompletions uses Protocols.OpenAI✕
- Connections
- 2 pairs of types, where a type in Protocols.ChatCompletions uses a type in Protocols.OpenAI.
- The types involved
RubyLLM.Protocols.ChatCompletions.Batches→RubyLLM.Protocols.OpenAI.BatchesRubyLLM.Protocols.ChatCompletions.EmbeddingBatches→RubyLLM.Protocols.OpenAI.Batches
- Direction
- Protocols.ChatCompletions uses Protocols.OpenAI, so a change to Protocols.OpenAI can break Protocols.ChatCompletions, while a change to Protocols.ChatCompletions can't break Protocols.OpenAI.
- Position
- Below the diagonal: Protocols.ChatCompletions uses a module listed above it, which is the healthy direction.
21→15 Protocols.Converse uses Protocols.Bedrock✕
- Connections
- 1 pair of types, where a type in Protocols.Converse uses a type in Protocols.Bedrock.
- The types involved
RubyLLM.Protocols.Converse.Batches→RubyLLM.Protocols.Bedrock.Batches
- Direction
- Protocols.Converse uses Protocols.Bedrock, so a change to Protocols.Bedrock can break Protocols.Converse, while a change to Protocols.Converse can't break Protocols.Bedrock.
- Position
- Below the diagonal: Protocols.Converse uses a module listed above it, which is the healthy direction.
22→17 Protocols.Interactions uses Protocols.Gemini✕
- Connections
- 1 pair of types, where a type in Protocols.Interactions uses a type in Protocols.Gemini.
- The types involved
RubyLLM.Protocols.Interactions.Transcription→RubyLLM.Protocols.Gemini.FileTranscription
- Direction
- Protocols.Interactions uses Protocols.Gemini, so a change to Protocols.Gemini can break Protocols.Interactions, while a change to Protocols.Interactions can't break Protocols.Gemini.
- Position
- Below the diagonal: Protocols.Interactions uses a module listed above it, which is the healthy direction.
23→18 Protocols.Responses uses Protocols.OpenAI✕
- Connections
- 1 pair of types, where a type in Protocols.Responses uses a type in Protocols.OpenAI.
- The types involved
RubyLLM.Protocols.Responses.Batches→RubyLLM.Protocols.OpenAI.Batches
- Direction
- Protocols.Responses uses Protocols.OpenAI, so a change to Protocols.OpenAI can break Protocols.Responses, while a change to Protocols.Responses can't break Protocols.OpenAI.
- Position
- Below the diagonal: Protocols.Responses uses a module listed above it, which is the healthy direction.
24→9 Protocols uses Protocols.Anthropic✕
- Connections
- 7 pairs of types, where a type in Protocols uses a type in Protocols.Anthropic.
- The types involved
RubyLLM.Protocols.Anthropic→RubyLLM.Protocols.Anthropic.ChatRubyLLM.Protocols.Anthropic→RubyLLM.Protocols.Anthropic.EmbeddingsRubyLLM.Protocols.Anthropic→RubyLLM.Protocols.Anthropic.MediaRubyLLM.Protocols.Anthropic→RubyLLM.Protocols.Anthropic.ModelsRubyLLM.Protocols.Anthropic→RubyLLM.Protocols.Anthropic.RequestShapesRubyLLM.Protocols.Anthropic→RubyLLM.Protocols.Anthropic.StreamingRubyLLM.Protocols.Anthropic→RubyLLM.Protocols.Anthropic.Tools
- Direction
- Protocols uses Protocols.Anthropic, so a change to Protocols.Anthropic can break Protocols, while a change to Protocols can't break Protocols.Anthropic.
- Position
- Below the diagonal: Protocols uses a module listed above it, which is the healthy direction.
24→10 Protocols uses Protocols.Cohere✕
- Connections
- 11 pairs of types, where a type in Protocols uses a type in Protocols.Cohere.
- The types involved
RubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.BatchesRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.ChatRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.EmbeddingsRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.MediaRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.ModelsRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.OCRRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.RerankRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.StreamingRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.TokenizationRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.ToolsRubyLLM.Protocols.Cohere→RubyLLM.Protocols.Cohere.Transcription
- Direction
- Protocols uses Protocols.Cohere, so a change to Protocols.Cohere can break Protocols, while a change to Protocols can't break Protocols.Cohere.
- Position
- Below the diagonal: Protocols uses a module listed above it, which is the healthy direction.
24→13 Protocols uses RubyLLM✕
- Connections
- 11 pairs of types, where a type in Protocols uses a type in RubyLLM.
- The types involved
RubyLLM.Protocols.Anthropic→RubyLLM.ProtocolRubyLLM.Protocols.ChatCompletions→RubyLLM.ProtocolRubyLLM.Protocols.Cohere→RubyLLM.ProtocolRubyLLM.Protocols.Converse→RubyLLM.ProtocolRubyLLM.Protocols.Deepgram→RubyLLM.ProtocolRubyLLM.Protocols.ElevenLabs→RubyLLM.ProtocolRubyLLM.Protocols.Files→RubyLLM.ProtocolRubyLLM.Protocols.Gemini→RubyLLM.ProtocolRubyLLM.Protocols.Interactions→RubyLLM.ProtocolRubyLLM.Protocols.InvokeModel→RubyLLM.ProtocolRubyLLM.Protocols.SystemOne→RubyLLM.Protocol
- Direction
- Protocols uses RubyLLM, so a change to RubyLLM can break Protocols, while a change to Protocols can't break RubyLLM.
- Position
- Below the diagonal: Protocols uses a module listed above it, which is the healthy direction.
24→16 Protocols uses Protocols.ElevenLabs✕
- Connections
- 4 pairs of types, where a type in Protocols uses a type in Protocols.ElevenLabs.
- The types involved
RubyLLM.Protocols.ElevenLabs→RubyLLM.Protocols.ElevenLabs.ModelsRubyLLM.Protocols.ElevenLabs→RubyLLM.Protocols.ElevenLabs.SpeechRubyLLM.Protocols.ElevenLabs→RubyLLM.Protocols.ElevenLabs.StreamingTranscriptionRubyLLM.Protocols.ElevenLabs→RubyLLM.Protocols.ElevenLabs.Transcription
- Direction
- Protocols uses Protocols.ElevenLabs, so a change to Protocols.ElevenLabs can break Protocols, while a change to Protocols can't break Protocols.ElevenLabs.
- Position
- Below the diagonal: Protocols uses a module listed above it, which is the healthy direction.
24→17 Protocols uses Protocols.Gemini✕
- Connections
- 13 pairs of types, where a type in Protocols uses a type in Protocols.Gemini.
- The types involved
RubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.CachesRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.ChatRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.EmbeddingsRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.ImagesRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.LongDecodeRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.MediaRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.ModelsRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.RequestShapesRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.SpeechRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.StreamingRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.ToolsRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.TranscriptionRubyLLM.Protocols.Gemini→RubyLLM.Protocols.Gemini.Videos
- Direction
- Protocols uses Protocols.Gemini, so a change to Protocols.Gemini can break Protocols, while a change to Protocols can't break Protocols.Gemini.
- Position
- Below the diagonal: Protocols uses a module listed above it, which is the healthy direction.
24→20 Protocols uses Protocols.ChatCompletions✕
- Connections
- 12 pairs of types, where a type in Protocols uses a type in Protocols.ChatCompletions.
- The types involved
RubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.ChatRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.EmbeddingsRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.ImagesRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.MediaRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.ModelsRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.ModerationRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.RequestShapesRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.SpeechRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.StreamingRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.ToolsRubyLLM.Protocols.ChatCompletions→RubyLLM.Protocols.ChatCompletions.TranscriptionRubyLLM.Protocols.Cohere→RubyLLM.Protocols.ChatCompletions.RequestShapes
- Direction
- Protocols uses Protocols.ChatCompletions, so a change to Protocols.ChatCompletions can break Protocols, while a change to Protocols can't break Protocols.ChatCompletions.
- Position
- Below the diagonal: Protocols uses a module listed above it, which is the healthy direction.
24→21 Protocols uses Protocols.Converse✕
- Connections
- 4 pairs of types, where a type in Protocols uses a type in Protocols.Converse.
- The types involved
RubyLLM.Protocols.Converse→RubyLLM.Protocols.Converse.ChatRubyLLM.Protocols.Converse→RubyLLM.Protocols.Converse.MediaRubyLLM.Protocols.Converse→RubyLLM.Protocols.Converse.RequestShapesRubyLLM.Protocols.Converse→RubyLLM.Protocols.Converse.Streaming
- Direction
- Protocols uses Protocols.Converse, so a change to Protocols.Converse can break Protocols, while a change to Protocols can't break Protocols.Converse.
- Position
- Below the diagonal: Protocols uses a module listed above it, which is the healthy direction.
24→22 Protocols uses Protocols.Interactions✕
- Connections
- 6 pairs of types, where a type in Protocols uses a type in Protocols.Interactions.
- The types involved
RubyLLM.Protocols.Interactions→RubyLLM.Protocols.Interactions.ChatRubyLLM.Protocols.Interactions→RubyLLM.Protocols.Interactions.ContentRubyLLM.Protocols.Interactions→RubyLLM.Protocols.Interactions.RequestShapesRubyLLM.Protocols.Interactions→RubyLLM.Protocols.Interactions.StreamingRubyLLM.Protocols.Interactions→RubyLLM.Protocols.Interactions.ToolsRubyLLM.Protocols.Interactions→RubyLLM.Protocols.Interactions.Transcription
- Direction
- Protocols uses Protocols.Interactions, so a change to Protocols.Interactions can break Protocols, while a change to Protocols can't break Protocols.Interactions.
- Position
- Below the diagonal: Protocols uses a module listed above it, which is the healthy direction.
24→23 Protocols uses Protocols.Responses✕
- Connections
- 6 pairs of types, where a type in Protocols uses a type in Protocols.Responses.
- The types involved
RubyLLM.Protocols.Responses→RubyLLM.Protocols.Responses.ApprovalsRubyLLM.Protocols.Responses→RubyLLM.Protocols.Responses.ChatRubyLLM.Protocols.Responses→RubyLLM.Protocols.Responses.MediaRubyLLM.Protocols.Responses→RubyLLM.Protocols.Responses.RequestShapesRubyLLM.Protocols.Responses→RubyLLM.Protocols.Responses.StreamingRubyLLM.Protocols.Responses→RubyLLM.Protocols.Responses.Tools
- Direction
- Protocols uses Protocols.Responses, so a change to Protocols.Responses can break Protocols, while a change to Protocols can't break Protocols.Responses.
- Position
- Below the diagonal: Protocols uses a module listed above it, which is the healthy direction.
25→24 Protocols.GPUStack uses Protocols✕
- Connections
- 1 pair of types, where a type in Protocols.GPUStack uses a type in Protocols.
- The types involved
RubyLLM.Protocols.GPUStack.Responses→RubyLLM.Protocols.Responses
- Direction
- Protocols.GPUStack uses Protocols, so a change to Protocols can break Protocols.GPUStack, while a change to Protocols.GPUStack can't break Protocols.
- Position
- Below the diagonal: Protocols.GPUStack uses a module listed above it, which is the healthy direction.
26→15 Protocols.InvokeModel uses Protocols.Bedrock✕
- Connections
- 1 pair of types, where a type in Protocols.InvokeModel uses a type in Protocols.Bedrock.
- The types involved
RubyLLM.Protocols.InvokeModel.EmbeddingBatches→RubyLLM.Protocols.Bedrock.Batches
- Direction
- Protocols.InvokeModel uses Protocols.Bedrock, so a change to Protocols.Bedrock can break Protocols.InvokeModel, while a change to Protocols.InvokeModel can't break Protocols.Bedrock.
- Position
- Below the diagonal: Protocols.InvokeModel uses a module listed above it, which is the healthy direction.
26→24 Protocols.InvokeModel uses Protocols✕
- Connections
- 5 pairs of types, where a type in Protocols.InvokeModel uses a type in Protocols.
- The types involved
RubyLLM.Protocols.InvokeModel.CohereEmbeddings→RubyLLM.Protocols.InvokeModelRubyLLM.Protocols.InvokeModel.NovaEmbeddings→RubyLLM.Protocols.InvokeModelRubyLLM.Protocols.InvokeModel.StabilityImages→RubyLLM.Protocols.InvokeModelRubyLLM.Protocols.InvokeModel.TitanMultimodalEmbeddings→RubyLLM.Protocols.InvokeModelRubyLLM.Protocols.InvokeModel.TitanTextEmbeddings→RubyLLM.Protocols.InvokeModel
- Direction
- Protocols.InvokeModel uses Protocols, so a change to Protocols can break Protocols.InvokeModel, while a change to Protocols.InvokeModel can't break Protocols.
- Position
- Below the diagonal: Protocols.InvokeModel uses a module listed above it, which is the healthy direction.
27→24 Protocols.Mistral uses Protocols✕
- Connections
- 2 pairs of types, where a type in Protocols.Mistral uses a type in Protocols.
- The types involved
RubyLLM.Protocols.Mistral.Conversations→RubyLLM.Protocols.ChatCompletionsRubyLLM.Protocols.Mistral.Files→RubyLLM.Protocols.Files
- Direction
- Protocols.Mistral uses Protocols, so a change to Protocols can break Protocols.Mistral, while a change to Protocols.Mistral can't break Protocols.
- Position
- Below the diagonal: Protocols.Mistral uses a module listed above it, which is the healthy direction.
28→1 Protocols.OpenRouter uses Batch✕
- Connections
- 1 pair of types, where a type in Protocols.OpenRouter uses a type in Batch.
- The types involved
RubyLLM.Protocols.OpenRouter.Batches→RubyLLM.Batch.Helpers
- Direction
- Protocols.OpenRouter uses Batch, so a change to Batch can break Protocols.OpenRouter, while a change to Protocols.OpenRouter can't break Batch.
- Position
- Below the diagonal: Protocols.OpenRouter uses a module listed above it, which is the healthy direction.
28→24 Protocols.OpenRouter uses Protocols✕
- Connections
- 2 pairs of types, where a type in Protocols.OpenRouter uses a type in Protocols.
- The types involved
RubyLLM.Protocols.OpenRouter.Files→RubyLLM.Protocols.FilesRubyLLM.Protocols.OpenRouter.Responses→RubyLLM.Protocols.Responses
- Direction
- Protocols.OpenRouter uses Protocols, so a change to Protocols can break Protocols.OpenRouter, while a change to Protocols.OpenRouter can't break Protocols.
- Position
- Below the diagonal: Protocols.OpenRouter uses a module listed above it, which is the healthy direction.
29→1 Protocols.VertexAI uses Batch✕
- Connections
- 1 pair of types, where a type in Protocols.VertexAI uses a type in Batch.
- The types involved
RubyLLM.Protocols.VertexAI.BatchPrediction→RubyLLM.Batch.Helpers
- Direction
- Protocols.VertexAI uses Batch, so a change to Batch can break Protocols.VertexAI, while a change to Protocols.VertexAI can't break Batch.
- Position
- Below the diagonal: Protocols.VertexAI uses a module listed above it, which is the healthy direction.
29→13 Protocols.VertexAI uses RubyLLM✕
- Connections
- 2 pairs of types, where a type in Protocols.VertexAI uses a type in RubyLLM.
- The types involved
RubyLLM.Protocols.VertexAI.EmbeddingPrediction→RubyLLM.ProtocolRubyLLM.Protocols.VertexAI.Ranking→RubyLLM.Protocol
- Direction
- Protocols.VertexAI uses RubyLLM, so a change to RubyLLM can break Protocols.VertexAI, while a change to Protocols.VertexAI can't break RubyLLM.
- Position
- Below the diagonal: Protocols.VertexAI uses a module listed above it, which is the healthy direction.
29→24 Protocols.VertexAI uses Protocols✕
- Connections
- 2 pairs of types, where a type in Protocols.VertexAI uses a type in Protocols.
- The types involved
RubyLLM.Protocols.VertexAI.Files→RubyLLM.Protocols.FilesRubyLLM.Protocols.VertexAI.Research→RubyLLM.Protocols.Interactions
- Direction
- Protocols.VertexAI uses Protocols, so a change to Protocols can break Protocols.VertexAI, while a change to Protocols.VertexAI can't break Protocols.
- Position
- Below the diagonal: Protocols.VertexAI uses a module listed above it, which is the healthy direction.
30→24 Protocols.XAI uses Protocols✕
- Connections
- 1 pair of types, where a type in Protocols.XAI uses a type in Protocols.
- The types involved
RubyLLM.Protocols.XAI.Files→RubyLLM.Protocols.Files
- Direction
- Protocols.XAI uses Protocols, so a change to Protocols can break Protocols.XAI, while a change to Protocols.XAI can't break Protocols.
- Position
- Below the diagonal: Protocols.XAI uses a module listed above it, which is the healthy direction.
31→23 Providers.Azure uses Protocols.Responses✕
- Connections
- 1 pair of types, where a type in Providers.Azure uses a type in Protocols.Responses.
- The types involved
RubyLLM.Providers.Azure.Responses→RubyLLM.Protocols.Responses.Compaction
- Direction
- Providers.Azure uses Protocols.Responses, so a change to Protocols.Responses can break Providers.Azure, while a change to Providers.Azure can't break Protocols.Responses.
- Position
- Below the diagonal: Providers.Azure uses a module listed above it, which is the healthy direction.
31→24 Providers.Azure uses Protocols✕
- Connections
- 3 pairs of types, where a type in Providers.Azure uses a type in Protocols.
- The types involved
RubyLLM.Providers.Azure.ChatCompletions→RubyLLM.Protocols.ChatCompletionsRubyLLM.Providers.Azure.Cohere→RubyLLM.Protocols.CohereRubyLLM.Providers.Azure.Responses→RubyLLM.Protocols.Responses
- Direction
- Providers.Azure uses Protocols, so a change to Protocols can break Providers.Azure, while a change to Providers.Azure can't break Protocols.
- Position
- Below the diagonal: Providers.Azure uses a module listed above it, which is the healthy direction.
32→4 Providers.Bedrock.Mantle uses Providers.Bedrock✕
- Connections
- 3 pairs of types, where a type in Providers.Bedrock.Mantle uses a type in Providers.Bedrock.
- The types involved
RubyLLM.Providers.Bedrock.Mantle.Anthropic→RubyLLM.Providers.Bedrock.MantleRubyLLM.Providers.Bedrock.Mantle.ChatCompletions→RubyLLM.Providers.Bedrock.MantleRubyLLM.Providers.Bedrock.Mantle.Responses→RubyLLM.Providers.Bedrock.Mantle
- Direction
- Providers.Bedrock.Mantle uses Providers.Bedrock, so a change to Providers.Bedrock can break Providers.Bedrock.Mantle, while a change to Providers.Bedrock.Mantle can't break Providers.Bedrock.
- Position
- Below the diagonal: Providers.Bedrock.Mantle uses a module listed above it, which is the healthy direction.
32→24 Providers.Bedrock.Mantle uses Protocols✕
- Connections
- 3 pairs of types, where a type in Providers.Bedrock.Mantle uses a type in Protocols.
- The types involved
RubyLLM.Providers.Bedrock.Mantle.Anthropic→RubyLLM.Protocols.AnthropicRubyLLM.Providers.Bedrock.Mantle.ChatCompletions→RubyLLM.Protocols.ChatCompletionsRubyLLM.Providers.Bedrock.Mantle.Responses→RubyLLM.Protocols.Responses
- Direction
- Providers.Bedrock.Mantle uses Protocols, so a change to Protocols can break Providers.Bedrock.Mantle, while a change to Providers.Bedrock.Mantle can't break Protocols.
- Position
- Below the diagonal: Providers.Bedrock.Mantle uses a module listed above it, which is the healthy direction.
33→23 Providers.Meta uses Protocols.Responses✕
- Connections
- 1 pair of types, where a type in Providers.Meta uses a type in Protocols.Responses.
- The types involved
RubyLLM.Providers.Meta.Responses→RubyLLM.Protocols.Responses.TokenCounting
- Direction
- Providers.Meta uses Protocols.Responses, so a change to Protocols.Responses can break Providers.Meta, while a change to Providers.Meta can't break Protocols.Responses.
- Position
- Below the diagonal: Providers.Meta uses a module listed above it, which is the healthy direction.
33→24 Providers.Meta uses Protocols✕
- Connections
- 1 pair of types, where a type in Providers.Meta uses a type in Protocols.
- The types involved
RubyLLM.Providers.Meta.Responses→RubyLLM.Protocols.Responses
- Direction
- Providers.Meta uses Protocols, so a change to Protocols can break Providers.Meta, while a change to Providers.Meta can't break Protocols.
- Position
- Below the diagonal: Providers.Meta uses a module listed above it, which is the healthy direction.
34→24 Providers.Ollama uses Protocols✕
- Connections
- 2 pairs of types, where a type in Providers.Ollama uses a type in Protocols.
- The types involved
RubyLLM.Providers.Ollama.ChatCompletions→RubyLLM.Protocols.ChatCompletionsRubyLLM.Providers.Ollama.SystemOne→RubyLLM.Protocols.SystemOne
- Direction
- Providers.Ollama uses Protocols, so a change to Protocols can break Providers.Ollama, while a change to Providers.Ollama can't break Protocols.
- Position
- Below the diagonal: Providers.Ollama uses a module listed above it, which is the healthy direction.
35→23 Providers.OpenAI uses Protocols.Responses✕
- Connections
- 2 pairs of types, where a type in Providers.OpenAI uses a type in Protocols.Responses.
- The types involved
RubyLLM.Providers.OpenAI.Responses→RubyLLM.Protocols.Responses.CompactionRubyLLM.Providers.OpenAI.Responses→RubyLLM.Protocols.Responses.TokenCounting
- Direction
- Providers.OpenAI uses Protocols.Responses, so a change to Protocols.Responses can break Providers.OpenAI, while a change to Providers.OpenAI can't break Protocols.Responses.
- Position
- Below the diagonal: Providers.OpenAI uses a module listed above it, which is the healthy direction.
35→24 Providers.OpenAI uses Protocols✕
- Connections
- 1 pair of types, where a type in Providers.OpenAI uses a type in Protocols.
- The types involved
RubyLLM.Providers.OpenAI.Responses→RubyLLM.Protocols.Responses
- Direction
- Providers.OpenAI uses Protocols, so a change to Protocols can break Providers.OpenAI, while a change to Providers.OpenAI can't break Protocols.
- Position
- Below the diagonal: Providers.OpenAI uses a module listed above it, which is the healthy direction.
36→20 Providers.GPUStack uses Protocols.ChatCompletions✕
- Connections
- 1 pair of types, where a type in Providers.GPUStack uses a type in Protocols.ChatCompletions.
- The types involved
RubyLLM.Providers.GPUStack.ChatCompletions→RubyLLM.Protocols.ChatCompletions.Rerank
- Direction
- Providers.GPUStack uses Protocols.ChatCompletions, so a change to Protocols.ChatCompletions can break Providers.GPUStack, while a change to Providers.GPUStack can't break Protocols.ChatCompletions.
- Position
- Below the diagonal: Providers.GPUStack uses a module listed above it, which is the healthy direction.
36→24 Providers.GPUStack uses Protocols✕
- Connections
- 1 pair of types, where a type in Providers.GPUStack uses a type in Protocols.
- The types involved
RubyLLM.Providers.GPUStack.ChatCompletions→RubyLLM.Protocols.ChatCompletions
- Direction
- Providers.GPUStack uses Protocols, so a change to Protocols can break Providers.GPUStack, while a change to Providers.GPUStack can't break Protocols.
- Position
- Below the diagonal: Providers.GPUStack uses a module listed above it, which is the healthy direction.
36→25 Providers.GPUStack uses Protocols.GPUStack✕
- Connections
- 2 pairs of types, where a type in Providers.GPUStack uses a type in Protocols.GPUStack.
- The types involved
RubyLLM.Providers.GPUStack.ChatCompletions→RubyLLM.Protocols.GPUStack.TokenizationRubyLLM.Providers.GPUStack.ChatCompletions→RubyLLM.Protocols.GPUStack.Videos
- Direction
- Providers.GPUStack uses Protocols.GPUStack, so a change to Protocols.GPUStack can break Providers.GPUStack, while a change to Providers.GPUStack can't break Protocols.GPUStack.
- Position
- Below the diagonal: Providers.GPUStack uses a module listed above it, which is the healthy direction.
37→24 Providers.Mistral uses Protocols✕
- Connections
- 1 pair of types, where a type in Providers.Mistral uses a type in Protocols.
- The types involved
RubyLLM.Providers.Mistral.ChatCompletions→RubyLLM.Protocols.ChatCompletions
- Direction
- Providers.Mistral uses Protocols, so a change to Protocols can break Providers.Mistral, while a change to Providers.Mistral can't break Protocols.
- Position
- Below the diagonal: Providers.Mistral uses a module listed above it, which is the healthy direction.
37→27 Providers.Mistral uses Protocols.Mistral✕
- Connections
- 2 pairs of types, where a type in Providers.Mistral uses a type in Protocols.Mistral.
- The types involved
RubyLLM.Providers.Mistral.ChatCompletions→RubyLLM.Protocols.Mistral.MultiCompletionRubyLLM.Providers.Mistral.Conversations→RubyLLM.Protocols.Mistral.Conversations
- Direction
- Providers.Mistral uses Protocols.Mistral, so a change to Protocols.Mistral can break Providers.Mistral, while a change to Providers.Mistral can't break Protocols.Mistral.
- Position
- Below the diagonal: Providers.Mistral uses a module listed above it, which is the healthy direction.
38→20 Providers.OpenRouter uses Protocols.ChatCompletions✕
- Connections
- 1 pair of types, where a type in Providers.OpenRouter uses a type in Protocols.ChatCompletions.
- The types involved
RubyLLM.Providers.OpenRouter.ChatCompletions→RubyLLM.Protocols.ChatCompletions.Rerank
- Direction
- Providers.OpenRouter uses Protocols.ChatCompletions, so a change to Protocols.ChatCompletions can break Providers.OpenRouter, while a change to Providers.OpenRouter can't break Protocols.ChatCompletions.
- Position
- Below the diagonal: Providers.OpenRouter uses a module listed above it, which is the healthy direction.
38→24 Providers.OpenRouter uses Protocols✕
- Connections
- 2 pairs of types, where a type in Providers.OpenRouter uses a type in Protocols.
- The types involved
RubyLLM.Providers.OpenRouter.ChatCompletions→RubyLLM.Protocols.ChatCompletionsRubyLLM.Providers.OpenRouter.SystemOne→RubyLLM.Protocols.SystemOne
- Direction
- Providers.OpenRouter uses Protocols, so a change to Protocols can break Providers.OpenRouter, while a change to Providers.OpenRouter can't break Protocols.
- Position
- Below the diagonal: Providers.OpenRouter uses a module listed above it, which is the healthy direction.
38→28 Providers.OpenRouter uses Protocols.OpenRouter✕
- Connections
- 1 pair of types, where a type in Providers.OpenRouter uses a type in Protocols.OpenRouter.
- The types involved
RubyLLM.Providers.OpenRouter.ChatCompletions→RubyLLM.Protocols.OpenRouter.Transcription
- Direction
- Providers.OpenRouter uses Protocols.OpenRouter, so a change to Protocols.OpenRouter can break Providers.OpenRouter, while a change to Providers.OpenRouter can't break Protocols.OpenRouter.
- Position
- Below the diagonal: Providers.OpenRouter uses a module listed above it, which is the healthy direction.
39→23 Providers.XAI uses Protocols.Responses✕
- Connections
- 1 pair of types, where a type in Providers.XAI uses a type in Protocols.Responses.
- The types involved
RubyLLM.Providers.XAI.Responses→RubyLLM.Protocols.Responses.Compaction
- Direction
- Providers.XAI uses Protocols.Responses, so a change to Protocols.Responses can break Providers.XAI, while a change to Providers.XAI can't break Protocols.Responses.
- Position
- Below the diagonal: Providers.XAI uses a module listed above it, which is the healthy direction.
39→24 Providers.XAI uses Protocols✕
- Connections
- 2 pairs of types, where a type in Providers.XAI uses a type in Protocols.
- The types involved
RubyLLM.Providers.XAI.ChatCompletions→RubyLLM.Protocols.ChatCompletionsRubyLLM.Providers.XAI.Responses→RubyLLM.Protocols.Responses
- Direction
- Providers.XAI uses Protocols, so a change to Protocols can break Providers.XAI, while a change to Providers.XAI can't break Protocols.
- Position
- Below the diagonal: Providers.XAI uses a module listed above it, which is the healthy direction.
39→30 Providers.XAI uses Protocols.XAI✕
- Connections
- 4 pairs of types, where a type in Providers.XAI uses a type in Protocols.XAI.
- The types involved
RubyLLM.Providers.XAI.ChatCompletions→RubyLLM.Protocols.XAI.StreamingTranscriptionRubyLLM.Providers.XAI.ChatCompletions→RubyLLM.Protocols.XAI.TokenizationRubyLLM.Providers.XAI.Responses→RubyLLM.Protocols.XAI.StreamingTranscriptionRubyLLM.Providers.XAI.Responses→RubyLLM.Protocols.XAI.Tokenization
- Direction
- Providers.XAI uses Protocols.XAI, so a change to Protocols.XAI can break Providers.XAI, while a change to Providers.XAI can't break Protocols.XAI.
- Position
- Below the diagonal: Providers.XAI uses a module listed above it, which is the healthy direction.
40→17 Providers.VertexAI uses Protocols.Gemini✕
- Connections
- 2 pairs of types, where a type in Providers.VertexAI uses a type in Protocols.Gemini.
- The types involved
RubyLLM.Providers.VertexAI.LiveTranscription→RubyLLM.Protocols.Gemini.LiveTranscriptionRubyLLM.Providers.VertexAI.Transcription→RubyLLM.Protocols.Gemini.FileTranscription
- Direction
- Providers.VertexAI uses Protocols.Gemini, so a change to Protocols.Gemini can break Providers.VertexAI, while a change to Providers.VertexAI can't break Protocols.Gemini.
- Position
- Below the diagonal: Providers.VertexAI uses a module listed above it, which is the healthy direction.
40→24 Providers.VertexAI uses Protocols✕
- Connections
- 5 pairs of types, where a type in Providers.VertexAI uses a type in Protocols.
- The types involved
RubyLLM.Providers.VertexAI.Anthropic→RubyLLM.Protocols.AnthropicRubyLLM.Providers.VertexAI.ChatCompletions→RubyLLM.Protocols.ChatCompletionsRubyLLM.Providers.VertexAI.EmbedContent→RubyLLM.Protocols.GeminiRubyLLM.Providers.VertexAI.Gemini→RubyLLM.Protocols.GeminiRubyLLM.Providers.VertexAI.Mistral→RubyLLM.Protocols.ChatCompletions
- Direction
- Providers.VertexAI uses Protocols, so a change to Protocols can break Providers.VertexAI, while a change to Providers.VertexAI can't break Protocols.
- Position
- Below the diagonal: Providers.VertexAI uses a module listed above it, which is the healthy direction.
40→37 Providers.VertexAI uses Providers.Mistral✕
- Connections
- 4 pairs of types, where a type in Providers.VertexAI uses a type in Providers.Mistral.
- The types involved
RubyLLM.Providers.VertexAI.Mistral→RubyLLM.Providers.Mistral.ChatRubyLLM.Providers.VertexAI.Mistral→RubyLLM.Providers.Mistral.EmbeddingsRubyLLM.Providers.VertexAI.Mistral→RubyLLM.Providers.Mistral.MediaRubyLLM.Providers.VertexAI.Mistral→RubyLLM.Providers.Mistral.Models
- Direction
- Providers.VertexAI uses Providers.Mistral, so a change to Providers.Mistral can break Providers.VertexAI, while a change to Providers.VertexAI can't break Providers.Mistral.
- Position
- Below the diagonal: Providers.VertexAI uses a module listed above it, which is the healthy direction.
At a glance — Code Health
At a glance — Architecture
At a glance — Maturity
At a glance — Readiness
At a glance — Security
At a glance — Domain Modelling
At a glance — Accessibility
At a glance — Performance
Security & Compliance — OWASP Top-10 mapping
| OWASP category | Findings | Severity |
|---|---|---|
| A05:2025 — Injection | 60 | High / Critical |
| A03:2025 — Software Supply Chain Failures | 5 | Medium |
Roadmap
First, strengthen the web-security posture by replacing weak cryptographic hashes with SHA-256 or non-cryptographic alternatives depending on their specific use case. Next, address the eight static analysis findings regarding secrets in workflow files, followed by the five dangerous execution issues in the Ruby source files. Finally, secure the supply chain by resolving the unverified third-party binary installation and establishing proper build provenance.
| Do this | Helps | Effort | Dimension |
|---|---|---|---|
| Resolve the 8 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED (4), REDACTED (3), REDACTED. | +3.3 pts | Low | Static Analysis (SAST) |
| Review each MD5/SHA1 use by what it protects: replace it with SHA-256+ (or a KDF for passwords) where the digest is security-relevant, and switch it to a non-cryptographic hash (xxhash, murmurhash) where it only derives an identifier such as a cache key. | +5.3 pts | Medium | Web-Security Posture |
| Resolve the 5 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED (3), REDACTED, REDACTED. | +2.3 pts | Low | Static Analysis (SAST) |
| Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing — start with REDACTED. | +2.3 pts | Low | Supply-chain Provenance & Signing |
| Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing. | +2.3 pts | Low | Supply-chain Provenance & Signing |
| Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing. | +2.3 pts | Low | Supply-chain Provenance & Signing |
| Record your significant design decisions. Write one short document per decision, giving its date, the situation, the decision and its consequences, and keep them together with your other design documents. A docs/adr folder with each file named by number and title is the easiest place to find them. | +1.8 pts | Medium | Architecture documentation |
| Explain in the README how to run the tests. Add a section called Testing to the README in the repository's top folder, saying how to run the test suite. | +1.3 pts | Medium | Documentation (README) |
File quality
| File | Score | Band | Worst signal |
|---|---|---|---|
| REDACTED | 2.5 | Slop | Static Analysis (SAST): High: REDACTED |
| REDACTED | 4.3 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 4.4 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 4.5 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 4.5 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 4.6 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 4.6 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 5.1 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 5.1 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 5.7 | Mixed | Static Analysis (SAST): Medium: REDACTED |
| REDACTED | 5.7 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 5.8 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 5.8 | Mixed | Static Analysis (SAST): High: REDACTED |
| REDACTED | 5.8 | Mixed | Static Analysis (SAST): High: REDACTED |
| lib/ruby_llm/mcp.rb | 6.3 | Mixed | God Classes: TooManyMethods: MCP |
| lib/ruby_llm/provider.rb | 6.3 | Mixed | God Classes: TooManyMethods: Provider |
| lib/ruby_llm/protocol.rb | 6.3 | Mixed | God Classes: TooManyMethods: Protocol |
| lib/ruby_llm/active_record/chat_methods.rb | 6.3 | Mixed | God Classes: TooManyMethods: ChatMethods |
| lib/generators/ruby_llm/provider/scaffold.rb | 7.0 | Mixed | God Classes: TooManyMethods: Scaffold |
| lib/ruby_llm/message.rb | 7.0 | Mixed | God Classes: TooManyMethods: Message |
How the grades work
A definite problem that already costs you something and drags the score down: a missing authorisation check, a dependency with a known exploit, a build that does not reproduce. Failure here tends to cause failures elsewhere.
Likely wrong, but not failing yet. It degrades the codebase over a longer horizon and can cause failures elsewhere — not urgent this week, not something to carry for two years either.
Recorded, with no effect on how the codebase functions. Present so the survey is complete, not because it needs doing.
Something this survey could not settle from the outside, and which could be critical or serious. Either a control was required and no positive evidence of it exists in the repository — a backup job that nothing shows was ever restored from proves nothing about restores — or our own analysis could not run over that part of the tree. This is not a clean result. These are excluded from the score rather than awarded a pass, so the number on the cover neither rewards nor penalises them: if you act on this survey without resolving them, you carry that risk yourself. Each one is named under Limitations.
Methodology & how to trust this report
Watchdog is a deep, periodic assessment — run each sprint, monthly, or quarterly, taking the time to go wider and deeper than a quick check and surfacing in one coherent report what you'd otherwise piece together from a dozen separate tools. It scores deterministically: the same commit yields the same score, every run. 66 of 66 evaluated dimensions are computed purely by tools and static analysis (confidence 1.0); none rely on LLM judgement. Overall confidence is 0.8 — the weighted average across measured dimensions; it falls as more of the score leans on LLM-assisted judgement and rises when it's fully tool-backed.
What we checked — 66 dimensions across the health lenses
- Can you open the finding? Real findings cite a repo-relative file and line you can open at the cited line — never an absolute scratch path. Here, 195 of 205 do; the remainder are repo-wide signals — a dimension-level measurement, not a single line.
- Is there a tool behind the number? Every score below names the method that produced it — Roslyn, git, a scanner, or (for a handful of documentation/naming dimensions) an LLM labelled sampled · advisory — not a narrative.
- Does re-running give the same result? Run it again on the same commit and the score — and this report, byte for byte — is identical. A report whose numbers move between runs is describing the run, not the code.
Tools & methods
| Method | Backs | Version | Evaluator |
|---|---|---|---|
| Roslyn static analysis | Complexity, cohesion, coupling, dead code, API surface, layering | 5.3.0 | ✓ deterministic |
| Native secret scanner | Hardcoded secrets / credentials | 1.0.0 | ✓ deterministic |
| Watchdog duplication detector (in-process) | Code duplication | 1.0.0 | ✓ deterministic |
| Coverage (coverlet / dotnet-coverage) | Line & branch coverage | 10.0.400 | ✓ deterministic |
| NuGet / dotnet | Outdated, vulnerable & deprecated dependencies | 10.0.400 | ✓ deterministic |
| git / LibGit2Sharp | Churn hotspots, knowledge concentration, history | 2.43.0 · 0.31.0 | ✓ deterministic |
| gitleaks · semgrep · trivy | Secrets in history, SAST, CVEs, IaC & container, PII / GDPR | 1.86.0 · 0.69.3 | ✓ deterministic |
| LLM (sampled · advisory) | Documentation quality, ADR conformance, naming — sampled over a bounded sample; advisory, never a deterministic measurement | — | ◐ LLM · sampled · advisory |
Run transparency — what happened this run
- D8 Code Coverage — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Coverage NOT READ here — but this repository measures it: a Codecov configuration (codecov.yml) and a coverage step in CI (`--coverage-report`) shows that coverage is collected and tracked in your own CI. The built-in collector has no runner for this ecosystem (.rb), so the analyzer could not read the number — a gap in the analyzer's language coverage, not an unmeasured repo. Not scored. To have the real number read, produce a coverage report in a standard format (SimpleCov — `require "simplecov"` in your `spec_helper.rb`/`test_helper.rb`, which writes `coverage/.resultset.json` (or lcov via `simplecov-lcov`)) and commit it — a hosted scan measures a clone of the repository, so a report that exists only in a working tree, a CI runner's or your own, never reaches it; the artefact is commonly gitignored, so `git add -f` that one file (or un-ignore its path) and commit it alongside the code it measures. You can widen what we reach: optional: produce a coverage report in a standard format (SimpleCov — `require "simplecov"` in your `spec_helper.rb`/`test_helper.rb`, which writes `coverage/.resultset.json` (or lcov via `simplecov-lcov`)) and commit it — a hosted scan measures a clone of the repository, so a report that exists only in a working tree, a CI runner's or your own, never reaches it; the artefact is commonly gitignored, so `git add -f` that one file (or un-ignore its path) and commit it alongside the code it measures — then the real number is read on the next scan.
- D16 Bus Factor — not measured this run — This is a true statement about the repository that carries nothing for its owner to act on, so it is reported here rather than as a defect in their code. Single-maintainer repository — bus factor is not applicable (145 contributor(s) across 2122 commit(s) sampled, automation and bot accounts excluded). One of them holds 82% of the history; the other 144 hold 0.1% each on average, below the 5% at which there is somebody to hand the work to. That is a single maintainer with drive-by contributors, not a team whose knowledge has concentrated — so the bus factor is not applicable and there is nothing here for the owner to act on.
- D21 Naming Consistency — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. D21's deterministic naming rules are gated on C# first (docs/plans/deterministic-engine.md, P5): each other language is switched on only after its own concept suite and A/B gate pass. Until then its declarations carry no naming verdict this run — this says nothing about whether its names are consistent.
- D22 Internal API Consistency — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. D22's deterministic call-shape rules are gated on C# first (docs/plans/deterministic-engine.md, P5): each other language is switched on only after its own concept suite and A/B gate pass. This says nothing about whether that surface is consistent.
- D24 Comment Value — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check's reader does not cover the language this repository's product is written in, so it had nothing of the product to read. That is a gap in this analyzer's language reach — not a finding about this repository.
- D27 Navigability — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Navigability is taken over call sites whose callee resolved. For C# that population includes calls into referenced assemblies; for this language only calls declared inside the repository count, so the abstention threshold reads the repository's use of its standard library as a resolution failure. That is a limit of this analyzer, not a build problem in this repository.
- D28 Secrets (history) — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. `gitleaks` TIMED OUT after 188.1s against a 188.1s budget, so we killed its process tree. The budget is ours; the work may simply be larger than it allows. Nothing about this repository is asserted by the absence of a result; the budget is ours to raise, and raising it is our backlog item, not an action for this repository's owner.
- D30 Dependency Vulnerabilities — measured, with a gap in what it reached — Watchdog measured this, but not all of it. What it did not reach is a gap on our side — a collector, parser or image we have not built yet — so the numbers on that dimension cover less than the repository, and the part left out is not evidence that it would have passed. The scanner produced no output at all, so no dependency was actually scanned. 1 of 2 declared ecosystem(s) (npm) WERE scanned and every vulnerability they reported is included in this result; osv was not, so this dimension's score covers less than the dependency surface this repository declares, and nothing here is evidence that osv is free of known-vulnerable dependencies.
- D32 Data Compliance (PII/GDPR) — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. `lib/ruby_llm/provider.rb` produced a parse error, so every rule in this engine's `gdpr.yml` was absent there. That absence is NOT a clean result: these rules detect personal data crossing a boundary into a log sink, a URL or browser storage, and a file that was never parsed cannot report any of the three. The rest of the tree analysed normally and its rows above stand; only these files are unaccounted for. You can widen what we reach: fix the syntax error (or exclude the file deliberately) and re-scan to cover it.
- D43 Malicious Dependencies — measured, with a gap in what it reached — Watchdog measured this, but not all of it. What it did not reach is a gap on our side — a collector, parser or image we have not built yet — so the numbers on that dimension cover less than the repository, and the part left out is not evidence that it would have passed. The scanner produced no output at all, so no dependency was actually scanned. 1 of 2 declared ecosystem(s) (npm) WERE scanned and every vulnerability they reported is included in this result; osv was not, so this dimension's score covers less than the dependency surface this repository declares, and nothing here is evidence that osv is free of known-vulnerable dependencies.
- AX1 Captive dependencies — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check reads Microsoft.Extensions.DependencyInjection registrations in C# and Spring beans in Java/Kotlin only, and this repository defines its own container in lib/ruby_llm/provider.rb, whose lifetimes are not modelled yet, so it had nothing of this repository's product to read. That is a gap in this analyzer's language reach — not a finding that the repository is free of what this check looks for.
- AXR1 Runtime accessibility — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Render strategy: boot via DevServer (dev). the dev server did not expose a crawlable HTTP endpoint in time — no runtime evidence.
- C1 Data Protection — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. 'ActiveRecord::Base' is imported (lib/ruby_llm/railtie.rb:41), but no entity, schema, table or model declaration this check reads was found beside it, so whether personal data is stored could not be decided.
- C3 Audit Trail — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. 'ActiveRecord::Base' is imported (lib/ruby_llm/railtie.rb:41), but no entity, schema, table or model declaration this check reads was found beside it, so whether personal data is stored could not be decided.
- C4 Data Retention — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. 'ActiveRecord::Base' is imported (lib/ruby_llm/railtie.rb:41), but no entity, schema, table or model declaration this check reads was found beside it, so whether personal data is stored could not be decided.
- C5 Data-Subject Rights — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. 'ActiveRecord::Base' is imported (lib/ruby_llm/railtie.rb:41), but no entity, schema, table or model declaration this check reads was found beside it, so whether personal data is stored could not be decided.
- M4 Documentation accuracy — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Documentation accuracy could not be assessed in this run — the check did not complete, so it is not scored. This is a gap in the analyzer, not a finding about this repository. Engine failure: FileNotFoundException: Could not find file 'spec/fixtures/vcr_cassettes/activerecord_actsas_attachment_handling_handles_attachments_in_ask_method.yml'.
- P8 Schema migrations — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. The schema may be created by a runner, a deploy script or a config file outside what this check reads, so an absence here is our blind spot rather than a missing migration strategy. The card abstains instead of scoring. You can widen what we reach: if the schema is created and evolved by a tool this check does not name, naming it lets us teach the detector to read it.
Limitations & what we did not check
Watchdog assesses the repository exactly as committed, and only the repository. By design it does not reach outside the source tree: the live cloud account, the running CI/CD pipeline, the host's branch-protection and approval rules, the production configuration, or a restore actually exercised against a backup are all out of scope. That boundary is a feature, not a gap — a repo-relative, deterministic scan re-runs identically on any commit and every finding opens at a real file and line, where a live audit can neither be reproduced nor traced. The visible consequence is that controls which leave no in-repo evidence are reported as "not evidenced" and excluded from the score rather than awarded a number a static scan cannot justify.
Per-dimension blind spots
- D1 Cyclomatic Complexity: Cyclomatic complexity counts branches statically — it cannot tell an essential decision tree from accidental tangle, nor see complexity that lives in data or configuration (large switch-case token tables, DSL lexers/parsers, data-as-code rule tables) rather than control flow: a tokenizer's many single-character cases read as high complexity though each branch is trivial.
- D2 Cognitive Complexity: Cognitive-complexity heuristics approximate how hard code is to follow; genuine domain difficulty and well-named intent that eases reading are not captured.
- D3 God Classes: "God class" is sized by members and responsibilities visible in the type — a deliberately broad facade over a coherent subsystem can read the same as an accidental grab-bag. For front-end JS the file-length check is cohesion-aware (a single-responsibility module — one class/IIFE — earns a 3× threshold), but cohesion is approximated from top-level declarations, not true dependency structure.
- D4 Code Duplication: Duplication is token-similarity — an in-process token-stream comparison over sliding windows, with type-aware normalization — so it finds copy-paste, not semantic duplication expressed differently. Committed machine-written code (scaffolded migrations, designer/codegen output, protobuf/OpenAPI stubs, model snapshots) is EXCLUDED — its repetition is the tool's, not the team's — so the score reflects hand-written duplication only.
- D6 Cohesion (LCOM4): LCOM4 cohesion is syntactic — it infers connectivity from which methods touch which fields/methods by name, not from real runtime behaviour or intent.
- D11 Test Reliability: Flakiness is inferred from history/markers — Watchdog runs the suite once (for coverage), not the repeated runs under varied conditions that reveal nondeterminism, so a flaky test never recorded as failing is invisible here.
- D12 Dependency Hygiene: Dependency health reads manifests and lockfiles — a vulnerability in a vendored/copied dependency, or risk from how a dependency is actually used, is outside this view.
- D13 Secret Scanning: Secret detection is signature- and entropy-based on the current tree — a secret that does not match a known pattern, or one already rotated, will not be flagged (a clean scan is "nothing matched", not "no secrets exist").
- D14 License Compliance: License compatibility is checked against declared package metadata and a policy — mislabelled or missing license metadata, and obligations that depend on how you distribute, are not resolved here.
- D15 Churn × Complexity Hotspots: Churn hotspots come from git history — a freshly imported or squashed repository has no churn signal, and recent rewrites can mask a historically risky file.
- D17 Explicit Debt: Acknowledged-debt signals (TODO/FIXME, suppressions, dead code) are textual — undocumented debt that nobody marked, and debt that lives in design rather than annotations, is invisible. Committed machine-written code (scaffolded migrations, designer/codegen output, generated stubs) is excluded — it is never the team's dead code to delete.
- D26 Project Cohesion: Project focus is sized from members/namespaces per project — a project that is broad by deliberate design reads the same as one that has sprawled.
- D29 Static Analysis (SAST): SAST findings are pattern-based (semgrep) — it finds classes of bug it has rules for; logic flaws, auth/authorization gaps and issues needing runtime context are out of reach (and clean means "no rule matched").
- D30 Dependency Vulnerabilities: CVE matching depends on accurate package/version metadata and on the advisory databases — a vulnerability with no published advisory, or in code not declared as a dependency, is not seen. Coverage needs a RESOLVED graph: an unpinned requirements.txt, or a pom without a resolved build, yields partial coverage rather than a clean verdict. An ecosystem the analyzer cannot scan is reported as unmeasured, never as clean.
- D34 Knowledge Freshness: Freshness is decayed commit RECENCY, not comprehension — code read often but rarely committed reads as orphaned, and stable code that genuinely needs no changes is penalised the same as forgotten code; bot/squash commits distort it like the bus factor.
- D35 Change Coupling: Change coupling is co-change in COMMITS — files split across separate commits, or coupled only through a shared config/build step, read as uncoupled, and a sweeping commit (rename/format) is excluded so it doesn't couple everything. It shows that files change together, not WHY: a high coupling can be a healthy cohesive pair as readily as a hidden leak.
- D43 Malicious Dependencies: Only packages some vulnerability database has already NAMED as malicious are seen — a compromise published in the last hours, or never reported at all, is invisible here, and this dimension reading 10 is not evidence that a dependency is trustworthy. There is no typosquat or dependency-confusion analysis: a package nobody has reported is simply absent from the feeds. Coverage is the dependency scan's: an ecosystem that could not be scanned is disclosed as unmeasured, never as clean.
- D44 Platform End-of-Life: The support table is FROZEN, so it goes out of date by losing RECALL: a release that ended support after the table was written is missed until the table is refreshed, and this dimension reading 10 is not evidence that a platform is current. Only platforms the repository DECLARES in a place this pass reads are seen — a runtime named only in a Dockerfile (D31's subject), in a CI workflow (D29's), or in a file this pass does not parse (go.mod, a Gemfile ruby directive) is invisible here, which is why a repository declaring none of them abstains rather than scoring. Only frameworks with a PUBLISHED support policy are tracked: React, Flask and Express publish none, so their age cannot be judged and their absence from a report is not a statement that they are supported.
- AC1 Text alternatives: Alt-text is detected structurally — the scan sees that an alternative EXISTS, not whether it meaningfully describes the image, and decorative-vs-missing is judged by attribute shape; runtime-injected images and a non-role=img decorative svg are out of scope. This is accessibility readiness, never a WCAG conformance claim.
- AC2 Forms & labels: Label association is read from static markup — a label wired up at runtime (JS-set aria-labelledby, framework-injected ids) reads as missing, a present label says nothing about whether its text is correct. A known UI-library field component (e.g. a JSX <TextField>) is now checked conservatively — flagged only when it carries NO label/aria-label/aria-labelledby/id/name — but wrapper/context-labelled libraries (Chakra/Radix FormControl+FormLabel) aren't statically visible (possible false positive) and non-JSX lowercased components are still skipped. A click handler on a plain element is now asked for a name too (it is a control the author declared), but the subtree test that answers it is deliberately generous: any DYNAMIC text expression in the subtree counts as a name, so an icon chosen by a ternary ({cond ? <IconA/> : <IconB/>}) reads as named, and a glyph component from a library the icon-import list does not know still names its parent. A clean result is "no unlabelled control found", not a labelling proof.
- AC3 Page structure: Page structure is read from the static markup tree — landmarks, headings and lang injected at runtime aren't seen, heading ORDER is checked structurally (not against the rendered visual hierarchy), and lang/title/main fire only on full documents, never partials, and the data-table check sees header-cell presence (a <th> exists), not whether each header correctly associates with its cells. Static readiness, not conformance.
- AC5 ARIA correctness: ARIA correctness is checked against the static role/attribute shape — roles/attributes set dynamically aren't seen, a valid role says nothing about whether it matches the element's real behaviour, and required-state checks are suppressed when a JSX spread could supply them. The two-branch toggle check (a control whose state is conveyed only by which of two mutually exclusive branches renders) reads CONDITIONALS THAT ARE ATTRIBUTES — Vue v-if/v-else/v-show and Alpine x-if/x-show — so the same toggle written as a Svelte {#if} block or a JSX ternary is control flow the markup model never projects as a branch and is not seen at all.
- AC6 Visual & motion safety: Contrast and motion safety are PARTIAL by construction — literal colours (hex/rgb/hsl/named) in inline styles, in-repo <style> blocks, in-repo .css files, var() tokens, Tailwind neutral utilities and CSS-in-JS top-level declarations are read (same-rule/same-element colour+background pairs only); computed/runtime/theme colour, external-CDN stylesheets, CSS-in-JS dynamic (${…}) and nested-selector colours, cross-element pairs and image contrast stay out of reach, so a clean result is bounded by what the static CSS itself shows.
- AC7 A11y enforcement: Enforcement is scored from in-repo config/CI evidence only — an a11y gate enforced in external tooling with no in-repo trace can't be credited, and a configured linter is presence, not proof the rules actually run or block a merge.
- AX10 Code composition: Role is inferred from namespace/folder convention, not semantics — a domain concept living in a folder named "Services" reads as application, and the split is lines-of-code, not business value. The business-logic-share score is a SOFT, FLOORED signal: it contributes to the Architecture lens but is floored at the Critical gate, so an infrastructure-heavy design (a gateway, an ETL, a driver) is legitimately low without being nuked to zero.
- DM6 Domain ↔ infrastructure boundary: Infrastructure reached through a hand-rolled wrapper, a domain-named facade, reflection, or a string-keyed service locator resolves to a non-infra type and isn't seen; the body scan is symbol resolution over syntax, not full dataflow. A clean result means "no resolved infra reference in a domain body", not a proof of purity.
- P4 Deployment & Rollback: Approval/branch-protection rules live in repository settings the scan cannot see — only their in-repo evidence (config files, workflows) is checked, so a control enforced purely in the host's settings reads as "not evidenced".
- P6 Release Hygiene: Rollback/observability controls are inferred from repo artefacts (pipelines, dashboards-as-code) — controls configured in external tooling, with no in-repo trace, cannot be credited.
Dimensions
D1 · Cyclomatic Complexity
3 method(s) exceeded the cyclomatic complexity threshold of 15; the worst was provider-coverage.update at 33. A further 1 method(s) were over the threshold but excluded as flat dispatchers (a long switch/match over independent cases: many branches, almost no nesting), the largest being Models.determine_model_family at 18 — they are counted neither in the figure above nor in this dimension's score. 1 file carries no cyclomatic complexity row at all for this reason — every one of its over-threshold methods was excluded, so the exclusion is disclosed nowhere in the file itself: lib/ruby_llm/providers/vertexai/models.rb (Models.determine_model_family at 18). They are named here because the per-file figures other dimensions report are taken BEFORE this exclusion, so such a file can show a high maximum complexity elsewhere in this report and nothing here, with nothing to reconcile the two.
What to do
- Bring the 3 bodies over 15 down to 15 or less in Cyclomatic Complexity — start with provider-coverage.update (cyclomatic 33), provider-coverage.root.addEventListener("click") (cyclomatic 20), Chat.render_payload (cyclomatic 17). — This score is capped by its worst body, so a finding fixed alone moves it by almost nothing — the next one down takes its place. Refactoring these 3 together lifts Cyclomatic Complexity from 8.9 to about 10.0/10, projected with the scoring formula itself and assuming each lands exactly at 15.
- Enforce Cyclomatic Complexity in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
D2 · Cognitive Complexity
3 method(s) exceeded the cognitive complexity threshold of 15; the worst was provider-coverage.update at 18.
What to do
- Bring the 3 bodies over 15 down to 15 or less in Cognitive Complexity — start with provider-coverage.update (cognitive 18), ResearchJob.research (cognitive 17), provider-coverage.root.addEventListener("click") (cognitive 16). — This score is capped by its worst body, so a finding fixed alone moves it by almost nothing — the next one down takes its place. Refactoring these 3 together lifts Cognitive Complexity from 9.6 to about 10.0/10, projected with the scoring formula itself and assuming each lands exactly at 15.
- Enforce Cognitive Complexity in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
D3 · God Classes
32 god class(es) detected.
What to do
- Resolve the 25 TooManyMethods finding(s) in God Classes — start with REDACTED (6), mcp.rb, provider.rb. — One of this dimension's main actionable groups (25 warning-level).
- Resolve the 7 FileTooLong finding(s) in God Classes — start with REDACTED, mcp.rb, models.rb. — One of this dimension's main actionable groups (7 warning-level).
- Enforce God Classes in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
D4 · Code Duplication
17 duplicated block group(s) detected.
What to do
- Resolve the 4 Duplicated block (9 lines × 2) finding(s) in Code Duplication — start with REDACTED, streaming.rb, nova_embeddings.rb. — One of this dimension's main actionable groups (4 warning-level).
- Resolve the 3 Duplicated block (5 lines × 2) finding(s) in Code Duplication — start with prompt.rb, tools.rb, decisions.rb. — One of this dimension's main actionable groups (3 warning-level).
- Resolve the 2 Duplicated block (8 lines × 2) finding(s) in Code Duplication — start with embedding.rb, REDACTED. — One of this dimension's main actionable groups (2 warning-level).
- Enforce Code Duplication in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Verified — provenance only; does not change the score.
D6 · Cohesion (LCOM4)
14 of 128 classes have LCOM4 above 3.
What to do
- Resolve the 14 Low cohesion finding(s) in Cohesion (LCOM4) — start with media.rb, install_generator.rb, chat_ui_generator.rb. — One of this dimension's main actionable groups (14 warning-level).
- Enforce Cohesion (LCOM4) in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Verified — provenance only; does not change the score.
D11 · Test Reliability
1 flaky across 1 measured tier(s). Ruby (RSpec: repository root, 460 spec files): measured (1 flaky).
What to do
- Resolve the 1 Flaky test finding(s) in Test Reliability. — One of this dimension's main actionable groups (1 issue-level).
D12 · Dependency Hygiene
0 outdated direct gem(s), 0 pinning defect(s). 69 declaration(s) graded against rubygems.org, none of them against a committed Gemfile.lock — this repository commits none, which a published gem correctly does not, so currency is read off the declared requirements: a declaration that does not cap below the current release resolves to it on `bundle install`. 1 declaration(s) carry a requirement that EXCLUDES the current release; that is a deliberate constraint (a supported-version ceiling or a held-back CI pin) and is counted, not charged. Whether any of these gems is DEPRECATED or YANKED is not graded — rubygems.org publishes no deprecation marker, and a yank removes a version rather than flagging it. Whether any is UNMAINTAINED is not graded either: no registry publishes a maintenance status, and release age does not stand in for one. Whether any is UNUSED is a source question, not a registry one. Known CVEs in this dependency graph are D30's question, read from the manifest there.
D13 · Secret Scanning
Secret scan ran and found no leaked secrets.
D14 · License Compliance
0 of 23 shipped gem(s) use a banned license. Licences were resolved from rubygems.org over the 23 runtime gem(s) this repository declares. It commits no Gemfile.lock to close that set over, so the verdict covers the declarations rather than the whole resolved graph. Development-group and `add_development_dependency` gems are excluded: they are not distributed with this repository. This repository publishes itself under MIT, which is its own choice and is not judged here.
D15 · Churn × Complexity Hotspots
Top hotspots: lib/ruby_llm/protocols/gemini/chat.rb (29×17=493); docs/assets/js/homepage.js (2×15=30)
What to do
- Resolve the 2 Hotspot finding(s) in Churn × Complexity Hotspots — start with REDACTED, homepage.js. — One of this dimension's main actionable groups (2 warning-level).
D17 · Explicit Debt
0 deducted task-comment markers across 49973 LoC (0.0/KLoC) → score 10.0. Task comments only: this repository's language is read without a compiler, so D17's suppression, dead-code and commented-out-code arms did not run and this score counts fewer marker kinds than a .NET repository's would.
D26 · Project Cohesion
1 of 2 build units (Bundler) flagged as possibly oversized/incoherent.
What to do
- Resolve the 1 Projects may be oversized for their cohesion finding(s) in Project Cohesion. — One of this dimension's main actionable groups (1 recommendation-level).
D29 · Static Analysis (SAST)
60 finding(s): 0 critical, 53 high, 7 medium, 0 low. 33 unpinned-GitHub-Actions row(s) are reported here but scored by D36 (supply-chain provenance), which measures that posture as `pinned_actions` — one pinning decision is charged once, not once per lens. semgrep hit a parse error in 8 file(s) — `docs/_includes/home.html`, `docs/_includes/jekyll_vitepress/head_end.html` (lines 4–5, lines 9–10, line 14), `docs/_includes/provider_coverage_matrix.html` (lines 1–3), `docs/_includes/version_select.html` (lines 1–11, line 14, lines 23–24, …), `docs/bin/build-api.sh` (line 6, line 9), … (+3 more) — so no absence of findings in the named regions is evidence of anything; rows reported elsewhere in those files are real. Fix the syntax error (or exclude the file deliberately) and re-scan to cover them. Separately, one or more rules could not re-parse an embedded snippet in 3 file(s) (e.g. a workflow `run:` block read as shell). Those files WERE scanned and their other rows are unaffected; only those rules' view of those snippets is missing.
What to do
- Resolve the 8 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED (4), REDACTED (3), REDACTED. — One of this dimension's main actionable groups (8 issue-level).
- Resolve the 5 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED (3), REDACTED, REDACTED. — One of this dimension's main actionable groups (5 issue-level).
- No action in Static Analysis (SAST) — all 32 REDACTED finding(s) are reported here at file:line but scored by D36 (supply-chain provenance), so none is charged to this dimension. — One of this dimension's main actionable groups (32 issue-level, 0 of them charged here).
D30 · Dependency Vulnerabilities
No known-vulnerable dependencies in the 1 ecosystem(s) that were scanned (npm). Partial dependency scan: 1 of 2 declared ecosystem(s) were scanned (npm), and the findings above are real and complete for them. osv was not scanned (osv: the scanner produced no output at all, so no dependency was actually scanned), so this is not the whole dependency surface and the result is reported at reduced confidence.
D34 · Knowledge Freshness
Every significant source file has living knowledge — recently and meaningfully worked. Counted over 247 of the 504 production source files in this repository: the rest are under the ~2,400-byte size floor this dimension measures over.
D35 · Change Coupling
Strongest change-coupling: embedding.rb↔image.rb 69%; embedding.rb↔transcription.rb 56%; image.rb↔transcription.rb 56%
What to do
- Resolve the 2 Change coupling finding(s) in Change Coupling — start with chat_methods.rb, context.rb. — One of this dimension's main actionable groups (2 warning-level).
- Resolve the 1 Change coupling clique finding(s) in Change Coupling — start with embedding.rb. — One of this dimension's main actionable groups (1 warning-level).
D36 · Supply-chain Provenance & Signing
0/4 supply-chain integrity signals present (provenance, signing, SBOM, pinned actions).
What to do
- Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing. — One of this dimension's main actionable groups (1 warning-level).
- Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing — start with REDACTED. — One of this dimension's main actionable groups (1 warning-level).
- Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing. — One of this dimension's main actionable groups (1 recommendation-level).
D43 · Malicious Dependencies
No known-vulnerable dependencies in the 1 ecosystem(s) that were scanned (npm). Partial dependency scan: 1 of 2 declared ecosystem(s) were scanned (npm), and the findings above are real and complete for them. osv was not scanned (osv: the scanner produced no output at all, so no dependency was actually scanned), so this is not the whole dependency surface and the result is reported at reduced confidence.
D44 · Platform End-of-Life
0 end-of-life runtime(s) and 0 end-of-life framework(s), read from 0 platform declaration(s) and 69 dependency declaration(s). This dimension reads what the repository says about ITSELF — a pinned target framework, a version file, a capped requires-python, a Rust toolchain pin, a framework major a constraint cannot move off. A FLOOR is deliberately never charged: `requires-python = ">=3.8"` states what the package SUPPORTS, not what it runs on, and a well-maintained library declares exactly that while running its own CI on a current release. The end-of-life facts are FROZEN and dated, so this dimension needs no network and answers identically inside a closed scan fence; as the table ages it loses recall and never precision, because a statement about support that ended in the past cannot become false. The OS layer of a container image is D31's question and the toolchain a CI workflow installs is D29's; this row is neither.
Frontend & cross-cutting dimensions
AC1 · Text alternatives
AC2 · Forms & labels
AC3 · Page structure
- No <main> (or role="main") means no "skip to content" target and a weaker landmark map. Wrap the primary content in <main>. (×3) — media/tour/video/hyperframes/index.html:2, media/tour/video/hyperframes/src/index.html:2, media/tour/video/hyperframes/src/partials/shell.html:2
What to do
- Declare <html lang>, a document <title> and a <main> landmark, keep headings in order, leave zoom enabled, title iframes and drop meta-refresh.
AC5 · ARIA correctness
AC6 · Visual & motion safety
- `.home-play-button` moves (`transition: transform 0.18s ease`) and nothing reads prefers-reduced-motion for it — no @media (prefers-reduced-motion) rule for this element, no reset this project loads — so a motion-sensitive user cannot turn it off (WCAG 2.3.3). Declare it under @media (prefers-reduced-motion: no-preference), or neutralise it under reduce. — docs/assets/css/rubyllm-overrides.css:678
- `.home-demo-soon` moves (`transition: opacity 0.18s ease, transform 0.18s ease`) and nothing reads prefers-reduced-motion for it — no @media (prefers-reduced-motion) rule for this element, no reset this project loads — so a motion-sensitive user cannot turn it off (WCAG 2.3.3). Declare it under @media (prefers-reduced-motion: no-preference), or neutralise it under reduce. — docs/assets/css/rubyllm-overrides.css:711
- `.home-model-switcher-model` moves (`transition: opacity 0.18s ease, transform 0.18s ease`) and nothing reads prefers-reduced-motion for it — no @media (prefers-reduced-motion) rule for this element, no reset this project loads — so a motion-sensitive user cannot turn it off (WCAG 2.3.3). Declare it under @media (prefers-reduced-motion: no-preference), or neutralise it under reduce. — docs/assets/css/rubyllm-overrides.css:1404
- `.home-love-nav` moves (`transition: color 0.18s ease, transform 0.18s ease`) and nothing reads prefers-reduced-motion for it — no @media (prefers-reduced-motion) rule for this element, no reset this project loads — so a motion-sensitive user cannot turn it off (WCAG 2.3.3). Declare it under @media (prefers-reduced-motion: no-preference), or neutralise it under reduce. — docs/assets/css/rubyllm-overrides.css:1559
- `.home-love-card` moves (`transition: box-shadow 0.18s ease, transform 0.18s ease`) and nothing reads prefers-reduced-motion for it — no @media (prefers-reduced-motion) rule for this element, no reset this project loads — so a motion-sensitive user cannot turn it off (WCAG 2.3.3). Declare it under @media (prefers-reduced-motion: no-preference), or neutralise it under reduce. — docs/assets/css/rubyllm-overrides.css:1609
What to do
- Keep a visible focus style (don't remove the outline without a replacement), guard animation with prefers-reduced-motion, and raise low-contrast colour pairs to at least 4.5:1.
AC7 · A11y enforcement
- No accessibility enforcement found — no a11y linter (an accessibility check that can read your UI — no component framework was detected and your pages are rendered by server-side templates, which neither the JSX/Vue ESLint plugins nor the HTML-template linters can parse; run axe/pa11y over the rendered pages, or assert the accessibility invariants over that rendered HTML in the test suite you already have) and no axe/pa11y/Lighthouse in tests or CI. Start by running that check over your rendered pages in CI. What was searched, so you can tell an absence from a miss: the 7 markup file(s) this pass actually assessed, the linter configuration checked in beside them, and this repository's test and CI files — matched by name against the accessibility checkers this dimension carries. An audit run outside the repository, a hosted scanner, or a check whose name is not one of those, is not seen here.
What to do
- Enforce accessibility in the toolchain: add an accessibility check that can read your UI — no component framework was detected and your pages are rendered by server-side templates, which neither the JSX/Vue ESLint plugins nor the HTML-template linters can parse; run axe/pa11y over the rendered pages, or assert the accessibility invariants over that rendered HTML in the test suite you already have, then assert with your test runner's axe binding (jest-axe, vitest-axe, cypress-axe or @axe-core/playwright) in tests, then gate axe/pa11y/Lighthouse in CI.
AX10 · Code composition
What to do
- The domain core is a small share of production code, but most of the rest matched no layer vocabulary at all — so this is not yet an anemic-domain finding. The namespace/path convention could not place that code, which makes the composition above a statement about the naming, not about the design. Name the layers (or check that the repository's conventions differ from the ones this check knows) before reading a thin domain into it.
AX2 · Stateful singletons
- `TokenEndpoint` serves requests (a Rack app is one instance the server shares across every request thread, so its instance variables are shared along with its class variables and globals) but updates that shared state with a read-modify-write outside any lock (self.class.requests; e.g. `self.class.requests` in `call` at line 26). Two requests that interleave between the read and the write lose one update. — benchmarks/vertex_ai.rb:17
What to do
- Ruby: keep class variables, globals and constants out of request handling — move counters and caches into a store built for concurrent access (the database, Rails.cache, Redis) or a Concurrent::AtomicFixnum / Concurrent::Map — or update them inside a shared Mutex's `synchronize` block; a Rack app's own instance variables are shared the same way.
AX3 · Project dependency cycles
AX4 · Dependency direction
AX6 · Interface segregation
AX8 · Test isolation
DM6 · Domain ↔ infrastructure boundary
GD1 · Unfinished & placeholder code
IC1 · Incompleteness & stubs
- A test is skipped/ignored — coverage that looks present but never runs. Re-enable it or delete it so the green run means something. (×9) — spec/ruby_llm/chat_bedrock_thinking_spec.rb:22, spec/ruby_llm/protocols/cohere/batches_spec.rb:258, spec/ruby_llm/protocols/cohere/batches_spec.rb:281, …
What to do
- Clear the softer debt: re-enable or delete `xit`/`skip`ped tests, decide whether each empty method or NotImplementedError refusal is deliberate (and say so in a comment), and replace `# rubocop:disable all` with targeted, justified disables.
M1 · Documentation (README)
What to do
- Explain in the README how to run the tests. Add a section called Testing to the README in the repository's top folder, saying how to run the test suite.
- Add a README to the 1 of 1 project(s) that lack one — worth up to 2 pts.
M2 · Architecture documentation
- No Architecture Decision Records found — no conventional ADR directory, no numbered `NNNN-title` documents in any markup this check reads, and nothing ADR-shaped by content. Design rationale recorded elsewhere (a design-notes tree, a mailing list, pull-request discussion) is not visible to this check and is not re-findable per decision, so a future maintainer cannot ask why one choice was made and get an answer.
What to do
- Record your significant design decisions. Write one short document per decision, giving its date, the situation, the decision and its consequences, and keep them together with your other design documents. A docs/adr folder with each file named by number and title is the easiest place to find them.
M3 · Folder & project structure
P1 · CI/CD gates
P10 · Library API & versioning
P3 · Security & performance tooling
P4 · Deployment & Rollback
P6 · Release Hygiene
P7 · Outbound HTTP resilience
- `Faraday.get` makes an outbound HTTP call, and nothing bounds it: no timeout, deadline, retry or circuit breaker is set for it here, and the client has no process-wide default. A slow or failing dependency will hold this service's request, thread or connection until the call gives up on its own — or never, for a client with no default timeout. — conformance/client.rb:40
What to do
- Set `open_timeout` and `read_timeout` explicitly on every client (Net::HTTP, Faraday `request: { timeout: }`, HTTParty `timeout:`), add Faraday's `:retry` middleware (faraday-retry) or `retriable`, and a breaker (`circuitbox`, `stoplight`) around dependencies that fail.
PF3 · Async & latency hygiene
S1 · Web-Security Posture
- MD5/SHA1 is computed here, and both are collision-broken. If this digest protects anything — a signature, an integrity or tamper check, a credential, or any value an attacker can influence — that is a real weakness: use SHA-256+ (`Digest::SHA256` or `OpenSSL::Digest.new("SHA256")`) for content integrity, or a KDF (bcrypt/Argon2/scrypt) for password storage. If it only derives a non-security identifier (a cache key, an avatar URL, an ETag), collision resistance carries no security consequence here; make that intent explicit instead — a non-cryptographic hash such as xxhash or murmurhash says it in code — since the algorithm alone cannot distinguish the two uses. Where a protocol this code does not own fixes the digest (a k-anonymity range query, a legacy signature format), the protocol decides; say so beside the call. — REDACTED:37
- `https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js` is executed by this page with no Subresource Integrity. Whoever can answer that request — the CDN, anyone who compromises it, anyone on the network path — runs arbitrary script in this page's origin, with its session. 3 such include(s) across the repository's markup. — media/tour/video/hyperframes/index.html:7
What to do
- Review each MD5/SHA1 use by what it protects: replace it with SHA-256+ (or a KDF for passwords) where the digest is security-relevant, and switch it to a non-cryptographic hash (xxhash, murmurhash) where it only derives an identifier such as a cache key.
- Pin third-party scripts and verify them: name an exact version in the URL and add an `integrity="sha384-…"` hash alongside `crossorigin="anonymous"` (both are required — an integrity hash on a cross-origin script without `crossorigin` is not evaluated, it blocks the script). Where the vendor ships a continuously-updated loader and publishes no stable hash (tag managers, analytics, chat widgets), Subresource Integrity is not available: constrain it instead with a `Content-Security-Policy` that names the exact origins allowed to execute, and drop the script from the pages that do not need it. Where the page is shipped inside a package that others host, prefer vendoring the asset and serving it from the app’s own origin, so no consumer inherits a third-party dependency they did not choose.
X1 · Async correctness
X12 · Unreachable branch
X13 · Undrained process stream
X14 · Bypassable address classification
X15 · Unvalidated length from an untrusted reader
X16 · Unfloored truncation loop
X17 · Uncapped recursion over a caller-supplied document
- `wrap` is public and takes `sources` — which it walks as a `Hash` — from its caller, and it calls itself at line 65 (`self.wrap`) with a piece of `sources` — and nowhere in `wrap` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends stack per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. A walk like this one raises `SystemStackError` past about 3,800 levels on the main thread and about 1,000 on a `Thread` — where Puma serves requests and Sidekiq runs jobs (MRI 3.3 and 4.0). `JSON.parse` stops at 100 levels by default, but the other parsers a caller may use do not: `YAML.safe_load` builds 1,400 levels from 2.8 KB, `Oj.load` 100,000, REXML 200,000, and `JSON.load` on json 2.7 (Ruby 3.3's default) or `JSON.parse(…, max_nesting: false)` 58,000. `SystemStackError` is not a `StandardError`: a bare `rescue` or `rescue => e` around the call does not stop it. Thread a depth argument through `wrap` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — lib/ruby_llm/attachment.rb:65
- `deep_merge` is public and takes `original` — which it walks as a `Hash` — from its caller, and it calls itself at line 121 (`self.deep_merge`) with a piece of `original` — and nowhere in `deep_merge` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends stack per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. A walk like this one raises `SystemStackError` past about 3,800 levels on the main thread and about 1,000 on a `Thread` — where Puma serves requests and Sidekiq runs jobs (MRI 3.3 and 4.0). `JSON.parse` stops at 100 levels by default, but the other parsers a caller may use do not: `YAML.safe_load` builds 1,400 levels from 2.8 KB, `Oj.load` 100,000, REXML 200,000, and `JSON.load` on json 2.7 (Ruby 3.3's default) or `JSON.parse(…, max_nesting: false)` 58,000. `SystemStackError` is not a `StandardError`: a bare `rescue` or `rescue => e` around the call does not stop it. Thread a depth argument through `deep_merge` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — lib/ruby_llm/support/utils.rb:121
- `deep_dup` is public and takes `value` — which it walks as a `Hash` — from its caller, and it calls itself at line 132 (`self.deep_dup`) with a piece of `value` (and at line 135) — and nowhere in `deep_dup` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends stack per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. A walk like this one raises `SystemStackError` past about 3,800 levels on the main thread and about 1,000 on a `Thread` — where Puma serves requests and Sidekiq runs jobs (MRI 3.3 and 4.0). `JSON.parse` stops at 100 levels by default, but the other parsers a caller may use do not: `YAML.safe_load` builds 1,400 levels from 2.8 KB, `Oj.load` 100,000, REXML 200,000, and `JSON.load` on json 2.7 (Ruby 3.3's default) or `JSON.parse(…, max_nesting: false)` 58,000. `SystemStackError` is not a `StandardError`: a bare `rescue` or `rescue => e` around the call does not stop it. Thread a depth argument through `deep_dup` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — lib/ruby_llm/support/utils.rb:132
- `deep_stringify_keys` is public and takes `value` — which it walks as a `Hash` — from its caller, and it calls itself at line 149 (`self.deep_stringify_keys`) with a piece of `value` (and at line 152) — and nowhere in `deep_stringify_keys` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends stack per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. A walk like this one raises `SystemStackError` past about 3,800 levels on the main thread and about 1,000 on a `Thread` — where Puma serves requests and Sidekiq runs jobs (MRI 3.3 and 4.0). `JSON.parse` stops at 100 levels by default, but the other parsers a caller may use do not: `YAML.safe_load` builds 1,400 levels from 2.8 KB, `Oj.load` 100,000, REXML 200,000, and `JSON.load` on json 2.7 (Ruby 3.3's default) or `JSON.parse(…, max_nesting: false)` 58,000. `SystemStackError` is not a `StandardError`: a bare `rescue` or `rescue => e` around the call does not stop it. Thread a depth argument through `deep_stringify_keys` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — lib/ruby_llm/support/utils.rb:149
- `deep_symbolize_keys` is public and takes `value` — which it walks as a `Hash` — from its caller, and it calls itself at line 165 (`self.deep_symbolize_keys`) with a piece of `value` (and at line 168) — and nowhere in `deep_symbolize_keys` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends stack per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. A walk like this one raises `SystemStackError` past about 3,800 levels on the main thread and about 1,000 on a `Thread` — where Puma serves requests and Sidekiq runs jobs (MRI 3.3 and 4.0). `JSON.parse` stops at 100 levels by default, but the other parsers a caller may use do not: `YAML.safe_load` builds 1,400 levels from 2.8 KB, `Oj.load` 100,000, REXML 200,000, and `JSON.load` on json 2.7 (Ruby 3.3's default) or `JSON.parse(…, max_nesting: false)` 58,000. `SystemStackError` is not a `StandardError`: a bare `rescue` or `rescue => e` around the call does not stop it. Thread a depth argument through `deep_symbolize_keys` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — lib/ruby_llm/support/utils.rb:165
- `references_only_definitions?` is public and takes `value` — which it walks as a `Hash` — from its caller, and it calls itself at line 47 (`self.references_only_definitions?`) with a piece of `value` (and at line 49) — and nowhere in `references_only_definitions?` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends stack per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. A walk like this one raises `SystemStackError` past about 3,800 levels on the main thread and about 1,000 on a `Thread` — where Puma serves requests and Sidekiq runs jobs (MRI 3.3 and 4.0). `JSON.parse` stops at 100 levels by default, but the other parsers a caller may use do not: `YAML.safe_load` builds 1,400 levels from 2.8 KB, `Oj.load` 100,000, REXML 200,000, and `JSON.load` on json 2.7 (Ruby 3.3's default) or `JSON.parse(…, max_nesting: false)` 58,000. `SystemStackError` is not a `StandardError`: a bare `rescue` or `rescue => e` around the call does not stop it. Thread a depth argument through `references_only_definitions?` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — lib/ruby_llm/protocols/gemini/schemas.rb:47
- `identified?` is public and takes `value` — which it walks as a `Hash` — from its caller, and it calls itself at line 65 (`self.identified?`) with a piece of `value` (and at line 66) — and nowhere in `identified?` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends stack per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. A walk like this one raises `SystemStackError` past about 3,800 levels on the main thread and about 1,000 on a `Thread` — where Puma serves requests and Sidekiq runs jobs (MRI 3.3 and 4.0). `JSON.parse` stops at 100 levels by default, but the other parsers a caller may use do not: `YAML.safe_load` builds 1,400 levels from 2.8 KB, `Oj.load` 100,000, REXML 200,000, and `JSON.load` on json 2.7 (Ruby 3.3's default) or `JSON.parse(…, max_nesting: false)` 58,000. `SystemStackError` is not a `StandardError`: a bare `rescue` or `rescue => e` around the call does not stop it. Thread a depth argument through `identified?` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — lib/ruby_llm/protocols/gemini/schemas.rb:65
- `copy` is public and takes `value` — which it walks as a `Hash` — from its caller, and it calls itself at line 12 (`self.copy`) with a piece of `value` — and nowhere in `copy` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends stack per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. A walk like this one raises `SystemStackError` past about 3,800 levels on the main thread and about 1,000 on a `Thread` — where Puma serves requests and Sidekiq runs jobs (MRI 3.3 and 4.0). `JSON.parse` stops at 100 levels by default, but the other parsers a caller may use do not: `YAML.safe_load` builds 1,400 levels from 2.8 KB, `Oj.load` 100,000, REXML 200,000, and `JSON.load` on json 2.7 (Ruby 3.3's default) or `JSON.parse(…, max_nesting: false)` 58,000. `SystemStackError` is not a `StandardError`: a bare `rescue` or `rescue => e` around the call does not stop it. Thread a depth argument through `copy` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — lib/ruby_llm/judge/data.rb:12
What to do
- Each finding names the externally-callable member that accepts the document and the self-call that descends, and states that no depth parameter or counter exists anywhere in the walk. Confirm from the two, then thread a depth argument through the walk and reject documents past a fixed maximum, or rewrite it with an explicit stack.
X18 · Disposal-pattern correctness
X19 · Unrestored process-global state
X21 · Side-effecting pattern guard
X22 · Contradicted release guard
X23 · Unguarded diagnostic materialisation
X24 · Document value interpolated into markup unescaped
X25 · Inert configuration knob
X26 · Unsynchronised callback handoff
X27 · Collection changed while being enumerated
X28 · Index access outside its own emptiness guard
X29 · Per-element action decided by a fixed element
X3 · Exception handling
- `rescue StandardError` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter. (×2) — lib/generators/ruby_llm/generator_helpers.rb:257, lib/generators/ruby_llm/generator_helpers.rb:361
What to do
- Swallowed exception (caught, then discarded)
X30 · Support guard that admits what it rejects
X5 · Nullable reference types
- This repository commits no Sorbet configuration (`sorbet/config`) and no `Steepfile`, so nothing type-checks its 491 Ruby product file(s): a nilable value reaches a method call silently and raises `NoMethodError: undefined method … for nil` the first time it is nil. A `# typed:` sigil does nothing without a `sorbet/config` beside the code; a Steep run configured only on a command line is not something this check can read.
What to do
- Adopt a checker and raise it file by file: with Sorbet, `srb init` (or `bundle exec tapioca init`) commits `sorbet/config`, then move each file's sigil from `# typed: false` to `# typed: true` and on to `strict` as its methods get `sig`s — `spoom bump` does the moving; with Steep, write RBS signatures and `check` the product directories in the `Steepfile` under `D::Ruby.default` or `strict`, not `lenient`. Then pay down the assertions rather than adding to them: each `T.must` is a nil the checker could not rule out — narrow it (`return unless x`, `x&.`, `if (y = x)`), make the attribute non-nilable and set it in `initialize`, or change the method so the nil cannot arrive.
WCAG coverage — what static analysis assessed
| Dimension | WCAG 2.2 A/AA criteria | Coverage |
|---|---|---|
| AC1 · Text alternatives | 1.1.1, 1.2.2, 1.2.5 | Partial signal |
| AC2 · Forms & labels | 1.3.1, 3.3.2, 4.1.2 | Partial signal |
| AC3 · Page structure | 1.4.4, 2.2.1, 2.4.1, 2.4.2, 3.1.1, 4.1.2 | Partial signal |
| AC5 · ARIA correctness | 4.1.2 | Partial signal |
| AC6 · Visual & motion safety | 1.4.3, 2.4.7 | Partial — literal CSS only |
| AC7 · A11y enforcement | enforcement — no page criterion | Enforcement posture (process) |
Reference — by lens
| Lens | Score | Rating | Impact |
|---|---|---|---|
| Code Health | 69% | Adequate — gated by X5 | Capped at Fair by a Critical contributor — resolve it before relying on this lens. |
| Architecture | 76% | Adequate — gated by D26 | Capped at Fair by a Critical contributor — resolve it before relying on this lens. |
| Maturity | 63% | Adequate — gated by M2 | Capped at Fair by a Critical contributor — resolve it before relying on this lens. |
| Readiness | 72% | Adequate — gated by P7 | Capped at Fair by a Critical contributor — resolve it before relying on this lens. |
| Security | 55% | Adequate — gated by D29, D36 | Capped at Fair by a Critical contributor — resolve it before relying on this lens. |
| Domain Modelling | 100% | Exemplary | Strongest area. |
| Accessibility | 67% | Adequate | Acceptable, with room to improve. |
| Performance | 100% | Exemplary | Solid. |
Unscored — 3 check(s) recorded observations but carry no score
- P12 CI test-gate honesty — 1 observation(s) recorded · Reported, not scored — this card publishes what the CI gate does with the test inventory rather than grading it. The findings above are its output.
- X10 Duplicated predicate — 25 observation(s) recorded · Advisory — this card reports evidence and never carries a score.
- X7 Silent fallback defaults — 4 observation(s) recorded · Advisory — this card reports evidence and never carries a score.
Not evidenced — 4 control(s) we could not find positive evidence for
- C3 Audit Trail — Not assessed: this repository stores data through 'ActiveRecord::Base' (lib/ruby_llm/railtie.rb:41) but declares no persisted model this check reads (ActiveRecord and Sequel migration columns, db/schema.rb, Mongoid document fields). Personal data may be written through raw SQL or a hosted backend, so whether its changes need an audit trail could not be established, and the trail's absence is not scored.
- C4 Data Retention — Not assessed: this repository stores data through 'ActiveRecord::Base' (lib/ruby_llm/railtie.rb:41) but declares no persisted model this check reads (ActiveRecord/Sequel migration and db/schema.rb columns, Mongoid document fields). Personal data may be written through raw SQL or a hosted backend, so whether a retention limit applies here could not be established, and its absence is not scored.
- C5 Data-Subject Rights — Not assessed: this repository stores data through 'ActiveRecord::Base' (lib/ruby_llm/railtie.rb:41) but declares no persisted model this check reads (ActiveRecord migration and db/schema.rb columns, Sequel migration columns, Mongoid fields). Personal data may be written through raw SQL or a hosted backend, so whether data-subject rights apply here could not be established, and their absence is not scored.
- P5 DR & Backup — not evidenced — repo shows no backup/RTO/RPO controls; absence of evidence is not evidence of a working control
Not included — 47 check(s) not relevant to this codebase
- AC4 Keyboard semantics — No interactive element found in the parsed markup — AC4 not applicable here.
- AX1 Captive dependencies — not analysed — these correctness checks read a source model that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
- AX5 Architecture & structure — not assessed — architecture style/structure is computed from a project graph (projects, types, module namespaces) that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
- AX7 Slice cohesion — not applicable — not a vertical-slice architecture
- AX9 CQS / query purity — no CQRS query handlers detected — query purity is not applicable to this codebase
- AXB2 Runtime readiness — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
- AXR1 Runtime accessibility — the dev server did not expose a crawlable HTTP endpoint in time — no runtime evidence This is a statement about this run, not a statement about your application: nothing here says the surface is inaccessible, only that it was never rendered.
- C1 Data Protection — Not assessed: this repository stores data through 'ActiveRecord::Base' (lib/ruby_llm/railtie.rb:41) but declares no persisted model this check reads (ActiveRecord migration and db/schema.rb columns, Sequel migration columns, Mongoid fields). Personal data may be written through raw SQL or a hosted backend, so the absence of a personal-data field here is not evidence that none is stored, and encryption at rest is not scored over a subject that could not be established.
- C2 Access Controls — No access-control surface detected in the analyzed source — this repository draws no route and defines no request handler (no Rails routes.draw or controller, no Sinatra, Roda or Grape application, no Hanami action). Access control is therefore N/A here — a gem is authorized by the application that mounts it, not by itself. If this codebase grows request handlers, the dimension reactivates and a default-deny posture is expected then.
- D10 Test Quality — ~70883 lines of test source are present (.rb) but no test cases reached the test census for this repository — no test root we could resolve declared them, or the files we read declare no cases in a test framework we recognise — so skipped/assertion-free tests couldn't be counted. Not scored — this is a gap in the analyzer, not a finding about this repository.
- D16 Bus Factor — single-maintainer repository — bus factor is not applicable
- D18 Solution Shape — D18 scores the shape of a .NET solution; this repository has no .NET solution or project files, so the dimension does not apply.
- D19 Documentation Quality — No runnable or published unit to look for an example in, and no multi-project root README to look for a map in; the other documentation kinds are withheld (see the gaps), so documentation quality is not applicable here.
- D20 ADR Quality — N/A — ADRs are expected on deployable products with a user-facing host, not consumed libraries; no ADR log is required here.
- D21 Naming Consistency — Naming consistency is measured on C# only; JavaScript, Python, Ruby, html not measured.
- D22 Internal API Consistency — API consistency is measured on C# only; the exposed Ruby surface was not measured.
- D23 Boundary Type-Coupling — No bounded-context organisation was detected either — neither a context-shaped layout, nor 2+ sibling source directories each declaring an aggregate root, nor 2+ projects each declaring an entity of their own. Aggregate roots were found in only one source directory (ruby_llm), which is a layered codebase, not two contexts. Declaring this codebase's bounded contexts (≥2) would let cross-boundary type coupling be assessed. Declare them in `.codehealth/config.yaml` at the repository root (create it if absent), mapping each context name to the module-path or namespace prefixes that belong to it — e.g. `architecture:` → `contexts:` → `Billing: ["src/billing", "Acme.Billing"]`, `Catalog: ["src/catalog", "Acme.Catalog"]`.
- D24 Comment Value — D24 reads C# and Visual Basic comments only — this repository's production source is .rb, which was left unread. Not scored: this is a gap in the analyzer, not a verdict about this repository.
- D25 ADR Conformance — no ADRs to check
- D27 Navigability — D27 counts a call into the standard library or a dependency as unresolved for this repository's language (.rb), so too few call sites counted as resolved to assess navigability. Not scored: this is a gap in the analyzer, not a verdict about this repository.
- D28 Secrets (history) — gitleaks killed before it finished — this repository's git history not scanned
- D31 IaC & Container Security — No Infrastructure-as-Code or container manifests found (Dockerfile, Docker Compose, Terraform, Kubernetes/Helm, CloudFormation, ARM, Bicep, Ansible); nothing to scan.
- D32 Data Compliance (PII/GDPR) — 1 file(s) were not parsed by semgrep — the PII/GDPR ruleset never ran over them
- D37 Vulnerability-disclosure Policy — No vulnerability-disclosure policy file found (SECURITY.md/.markdown/.rst/.txt at root or under .github/.forgejo/.gitea/docs, .well-known/security.txt). A coordinated-disclosure policy may live off-repo, so this is not evidenced rather than failed.
- D39 IL Efficiency — D39 measures the IL emitted by a .NET build; this repository has no .NET solution or project files, so the dimension does not apply.
- D40 Network Egress Confinement — No Kubernetes/orchestration workloads found in the repository manifests; network egress policy is a cluster-native control that may live at the platform/firewall layer, so there is nothing to assess here.
- D41 Kernel & Syscall Confinement — No Kubernetes/orchestration workloads found in the repository manifests; seccomp/AppArmor/SELinux confinement is a workload-level control, so there is nothing to assess here.
- D42 Runtime Threat Enforcement — No Kubernetes/orchestration workloads found in the repository manifests; runtime threat-detection and admission-control policy are cluster-level controls, so there is nothing to assess here.
- D5 Coupling — Not applicable — this RubyGems build ships 1 production module(s), so there is no coupling BETWEEN modules to measure. (Its test and non-production modules are not part of the shipped graph.)
- D7 Architectural Integrity — no checkable ADRs, and no project-reference graph for the cycle pass to read — so this dimension makes no claim about dependency cycles in either direction (where this repository's language has an import-cycle lens, cycles are reported there). Architectural integrity not assessed
- D8 Code Coverage — Coverage not included — suite not readable by the collector
- D9 Test Distribution — Test source is present (.rb) but no test cases reached the test census for this repository — no test root we could resolve declared them, or the files we read declare no cases in a test framework we recognise — so its unit/integration/BDD/E2E split couldn't be counted. Not scored — this is a gap in the analyzer, not a finding about this repository.
- ED1 Event-Driven — not scored — this repository shows none of the 3 signals this lens looks for
- ED5 Idempotency — no mutating command handlers or message consumers detected — idempotency check not applicable
- ES1 Event Sourcing — not scored — this repository shows none of the 3 signals this lens looks for
- M4 Documentation accuracy — Documentation accuracy could not be assessed in this run — the check did not complete, so it is not scored. This is a gap in the analyzer, not a finding about this repository.
- P2 Observability — This repository's JavaScript/TypeScript, Ruby source (3 module(s), 499 file(s) read) declares no entry point and bootstraps no server, and nothing here deploys a service — it is a library, run inside whatever hosts it, so production observability (structured logging, tracing/metrics, health checks) is N/A. If it grows a binary or a service, the dimension reactivates.
- P8 Schema migrations — ActiveRecord is declared, but the schema-migration mechanism could not be identified. Our limit, not a verdict on this repository.
- P9 Domain vs controller coverage — no coverage report found on disk — produce a coverage report in a standard format (SimpleCov — `require "simplecov"` in your `spec_helper.rb`/`test_helper.rb`, which writes `coverage/.resultset.json` (or lcov via `simplecov-lcov`)) and commit it — a hosted scan measures a clone of the repository, so a report that exists only in a working tree, a CI runner's or your own, never reaches it; the artefact is commonly gitignored, so `git add -f` that one file (or un-ignore its path) and commit it alongside the code it measures, or wire coverage collection into CI, to enable this cross-layer check
- PF1 Benchmark discipline — Not applicable: no benchmark suite was found. This check searched for `x.report(…)` entries in a `Benchmark.ips`/`bm`/`bmbm`/`benchmark`/`memory` block, `benchmark:` entries in a benchmark-driver YAML under a bench* directory, rspec-benchmark `perform_*` expectations where the gem is installed, or benchmark-ips, benchmark-memory, rspec-benchmark or benchmark-driver in a Gemfile or gemspec, and for a `*benchmark*` script that this repository's CI runs. Benchmarks are credited as a bonus, so their absence is neither scored nor deducted.
- PF2 Allocation hygiene — Not applicable: Ruby runs on a garbage-collected runtime that gives a program no allocation-control idiom to choose on a hot path — no pools, stack allocation or value types — so allocation awareness is not something this code can be rated on.
- X2 Cancellation propagation — no C# was loaded; this repository's Ruby was read, and no request in it can be cancelled by a caller — a Ruby request takes no token: Net::HTTP, Faraday and async-http are stopped from outside the call (Async::Task#stop, Thread#raise, Timeout), not through a parameter
- X20 Mistyped argument guard — This check asks whether an argument guard reports an EMPTY value as a NULL one, which needs a language that throws a null-specific argument exception — .NET's ArgumentNullException, the JVM's NullPointerException, Dart's ArgumentError.notNull. This repository contains none of those languages: the ones it is written in have a single exception for a bad argument, so there is no pair of exceptions to confuse and nothing here for it to assess. Not a gap in the analyzer and not a finding about your code.
- X32 Type resolved by simple name across every loaded assembly — This check is about how a .NET program searches the assemblies loaded into its process for a type, and this repository contains no .NET source, so there is nothing here for it to assess. Not a gap in the analyzer and not a finding about your code.
- X4 Structured logging — This check asks whether log calls keep a message's values apart from its text, which needs a logger that can carry them as fields. This repository's Ruby was read, and it logs through none: the standard library's Logger, which Rails.logger is, takes one finished message and no template arguments or fields (as log4net does in C#), so an interpolated message there loses nothing the API could have kept. The console gem (`Console.info`), SemanticLogger, Ougai and dry-logger would be judged.
- X6 Hand-rolled structured-format parsing — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
- X9 Subsumed condition operand — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
Appendix A — Findings (grouped)
Critical — 54 finding(s)
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- Flaky test: RubyLLM::Chat provider mapping lets provider options override the mapped value
- REDACTED
- REDACTED
- REDACTED
- REDACTED
Serious — 107 finding(s)
- TooManyMethods: Chat REDACTED:25
- TooManyMethods: MCP lib/ruby_llm/mcp.rb:32
- TooManyMethods: Provider lib/ruby_llm/provider.rb:35
- TooManyMethods: Protocol lib/ruby_llm/protocol.rb:37
- TooManyMethods: Agent lib/ruby_llm/agent.rb:42
- TooManyMethods: ChatMethods lib/ruby_llm/active_record/chat_methods.rb:26
- TooManyMethods: Models lib/ruby_llm/models.rb:17
- TooManyMethods: CodeExecution lib/ruby_llm/code_execution.rb:25
- TooManyMethods: Chat lib/ruby_llm/protocols/converse/chat.rb:9
- TooManyMethods: Chat lib/ruby_llm/protocols/chat_completions/chat.rb:7
- TooManyMethods: Chat lib/ruby_llm/protocols/anthropic/chat.rb:7
- TooManyMethods: Chat lib/ruby_llm/protocols/responses/chat.rb:7
- TooManyMethods: Attachment lib/ruby_llm/attachment.rb:19
- TooManyMethods: Cost lib/ruby_llm/cost.rb:28
- TooManyMethods: GeneratorHelpers lib/generators/ruby_llm/generator_helpers.rb:6
- TooManyMethods: Streaming lib/ruby_llm/protocols/converse/streaming.rb:11
- TooManyMethods: Bedrock lib/ruby_llm/providers/bedrock.rb:6
- TooManyMethods: Tool lib/ruby_llm/tool.rb:40
- TooManyMethods: Chat lib/ruby_llm/protocols/gemini/chat.rb:7
- TooManyMethods: Scaffold lib/generators/ruby_llm/provider/scaffold.rb:12
- TooManyMethods: Batch lib/ruby_llm/batch.rb:18
- TooManyMethods: AttachmentHelpers REDACTED:7
- TooManyMethods: RubyLLM lib/ruby_llm.rb:230
- TooManyMethods: Message lib/ruby_llm/message.rb:17
- TooManyMethods: MessageMethods lib/ruby_llm/active_record/message_methods.rb:11
- Low cohesion: Gemini (LCOM4 20) lib/ruby_llm/protocols/gemini/media.rb:8
- Low cohesion: InstallGenerator (LCOM4 13) lib/generators/ruby_llm/install/install_generator.rb:10
- Low cohesion: ChatUIGenerator (LCOM4 12) lib/generators/ruby_llm/chat_ui/chat_ui_generator.rb:9
- Low cohesion: ToolGenerator (LCOM4 11) lib/generators/ruby_llm/tool/tool_generator.rb:9
- Low cohesion: UpgradeGenerator (LCOM4 10) lib/generators/ruby_llm/upgrade/upgrade_generator.rb:10
- Low cohesion: Scaffold (LCOM4 7) lib/generators/ruby_llm/provider/scaffold.rb:12
- Low cohesion: Message (LCOM4 7) lib/ruby_llm/message.rb:17
- Low cohesion: Protocol (LCOM4 7) lib/ruby_llm/protocol.rb:37
- Low cohesion: Cohere (LCOM4 6) lib/ruby_llm/protocols/cohere.rb:7
- Low cohesion: MCP (LCOM4 4) lib/ruby_llm/mcp.rb:32
- Low cohesion: Task (LCOM4 4) lib/ruby_llm/mcp/task.rb:21
- Low cohesion: Model (LCOM4 4) lib/ruby_llm/model.rb:14
- Low cohesion: Provider (LCOM4 4) lib/ruby_llm/provider.rb:35
- Low cohesion: ModelCatalogPage (LCOM4 4) tasks/support/model_catalog_page.rb:7
- Recursive walk over a caller-supplied document has no depth cap lib/ruby_llm/attachment.rb:65
- Recursive walk over a caller-supplied document has no depth cap lib/ruby_llm/support/utils.rb:121
- Recursive walk over a caller-supplied document has no depth cap lib/ruby_llm/support/utils.rb:132
- Recursive walk over a caller-supplied document has no depth cap lib/ruby_llm/support/utils.rb:149
- Recursive walk over a caller-supplied document has no depth cap lib/ruby_llm/support/utils.rb:165
- Recursive walk over a caller-supplied document has no depth cap lib/ruby_llm/protocols/gemini/schemas.rb:47
- Recursive walk over a caller-supplied document has no depth cap lib/ruby_llm/protocols/gemini/schemas.rb:65
- Recursive walk over a caller-supplied document has no depth cap lib/ruby_llm/judge/data.rb:12
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- FileTooLong: REDACTED REDACTED
- FileTooLong: ruby_llm/mcp.rb lib/ruby_llm/mcp.rb
- FileTooLong: ruby_llm/models.rb lib/ruby_llm/models.rb
- FileTooLong: ruby_llm/protocol.rb lib/ruby_llm/protocol.rb
- FileTooLong: ruby_llm/agent.rb lib/ruby_llm/agent.rb
- FileTooLong: active_record/chat_methods.rb lib/ruby_llm/active_record/chat_methods.rb
- FileTooLong: ruby_llm/provider.rb lib/ruby_llm/provider.rb
- Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css:678
- Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css:711
- Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css:1404
- Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css:1559
- Animation without a prefers-reduced-motion guard docs/assets/css/rubyllm-overrides.css:1609
- Duplicated block (9 lines × 2) REDACTED:125
- Duplicated block (9 lines × 2) lib/ruby_llm/protocols/chat_completions/streaming.rb:40
- Duplicated block (9 lines × 2) lib/ruby_llm/protocols/invoke_model/nova_embeddings.rb:12
- Duplicated block (9 lines × 2) lib/ruby_llm/protocols/anthropic/files.rb:25
- Page without a main landmark media/tour/video/hyperframes/index.html:2
- Page without a main landmark media/tour/video/hyperframes/src/index.html:2
- Page without a main landmark media/tour/video/hyperframes/src/partials/shell.html:2
- Duplicated block (5 lines × 2) lib/ruby_llm/prompt.rb:68
- Duplicated block (5 lines × 2) lib/ruby_llm/protocols/cohere/tools.rb:124
- Duplicated block (5 lines × 2) lib/ruby_llm/protocols/openai/decisions.rb:98
- Hotspot: lib/ruby_llm/protocols/gemini/chat.rb lib/ruby_llm/protocols/gemini/chat.rb:34
- Hotspot: docs/assets/js/homepage.js docs/assets/js/homepage.js:123
- Change coupling: chat_methods.rb ↔ message_methods.rb lib/ruby_llm/active_record/chat_methods.rb:1
- Change coupling: context.rb ↔ provider.rb lib/ruby_llm/context.rb:1
- Duplicated block (8 lines × 2) lib/ruby_llm/embedding.rb:133
- Duplicated block (8 lines × 2) lib/ruby_llm/protocols/converse/chat.rb:386
- Duplicated block (6 lines × 2) lib/ruby_llm/protocols/anthropic/chat.rb:447
- Duplicated block (6 lines × 2) lib/ruby_llm/protocols/chat_completions/chat.rb:282
- Duplicated block (7 lines × 2) lib/ruby_llm/protocols/chat_completions/chat.rb:272
- Duplicated block (7 lines × 2) lib/ruby_llm/protocols/bedrock/batches.rb:80
- Swallowed exception (caught, then discarded) lib/generators/ruby_llm/generator_helpers.rb:257
- Swallowed exception (caught, then discarded) lib/generators/ruby_llm/generator_helpers.rb:361
- Accessibility enforcement below the top rung
- Stateful handler: TokenEndpoint benchmarks/vertex_ai.rb:17
- provider-coverage.update (cyclomatic 33) docs/assets/js/provider-coverage.js:27
- provider-coverage.root.addEventListener("click") (cyclomatic 20) docs/assets/js/provider-coverage.js:127
- Chat.render_payload (cyclomatic 17) lib/ruby_llm/protocols/gemini/chat.rb:34
- provider-coverage.update (cognitive 18) docs/assets/js/provider-coverage.js:27
- ResearchJob.research (cognitive 17) lib/ruby_llm/research_job.rb:92
- provider-coverage.root.addEventListener("click") (cognitive 16) docs/assets/js/provider-coverage.js:127
- Change coupling clique: embedding.rb, image.rb, transcription.rb lib/ruby_llm/embedding.rb
- REDACTED
- REDACTED
- Duplicated block (13–14 lines × 2) lib/ruby_llm/image.rb:76
- Duplicated block (13 lines × 2) lib/ruby_llm/ocr.rb:64
- Duplicated block (24 lines × 2) lib/ruby_llm/providers/hetzner/chat.rb:11
- Duplicated block (9 lines × 3) lib/ruby_llm/file_store/tools/search.rb:24
- Coverage collected but not gated
- Outbound HTTP without resilience conformance/client.rb:40
- Weak hash algorithm REDACTED:37
- Third-party script without Subresource Integrity media/tour/video/hyperframes/index.html:7
Minor — 44 finding(s)
- Duplicated predicate lib/ruby_llm/providers/bedrock.rb:99
- Duplicated predicate lib/ruby_llm/protocols/elevenlabs/flows/media.rb:9
- Duplicated predicate lib/ruby_llm/protocols/gemini/media.rb:55
- Duplicated predicate lib/ruby_llm/protocols/converse/media.rb:160
- Duplicated predicate lib/ruby_llm/providers/hetzner/chat.rb:22
- Duplicated predicate lib/ruby_llm/protocols/anthropic/chat.rb:434
- Duplicated predicate lib/ruby_llm/protocols/responses/chat.rb:372
- Duplicated predicate lib/ruby_llm/protocols/anthropic/chat.rb:41
- Duplicated predicate lib/ruby_llm/file_store/tools/edit.rb:53
- Duplicated predicate lib/ruby_llm/protocols/bedrock/files.rb:21
- Duplicated predicate lib/ruby_llm/protocols/gemini/chat.rb:189
- Duplicated predicate lib/ruby_llm/protocols/gemini/chat.rb:213
- Duplicated predicate lib/ruby_llm/prompt.rb:68
- Duplicated predicate lib/ruby_llm/protocols/deepgram/transcription.rb:34
- Duplicated predicate lib/ruby_llm/protocols/bedrock/rerank.rb:67
- Duplicated predicate lib/ruby_llm/protocols/anthropic/chat.rb:440
- Duplicated predicate lib/ruby_llm/providers/vertexai/anthropic/batches.rb:45
- Duplicated predicate lib/ruby_llm/protocols/converse/chat.rb:240
- Duplicated predicate lib/ruby_llm/protocols/chat_completions/chat.rb:282
- Duplicated predicate lib/ruby_llm/protocols/gemini/request_shapes.rb:48
- Duplicated predicate lib/ruby_llm/file_store/tools/list.rb:33
- Duplicated predicate lib/ruby_llm/protocols/chat_completions/embedding_batches.rb:26
- Duplicated predicate lib/ruby_llm/protocols/chat_completions/batches.rb:23
- Duplicated predicate lib/ruby_llm/providers/azure/videos.rb:54
- Duplicated predicate lib/ruby_llm/protocols/anthropic/tools.rb:61
- Disabled test spec/ruby_llm/chat_bedrock_thinking_spec.rb:22
- Disabled test spec/ruby_llm/protocols/cohere/batches_spec.rb:258
- Disabled test spec/ruby_llm/protocols/cohere/batches_spec.rb:281
- Disabled test spec/ruby_llm/protocols/perplexity/router_spec.rb:133
- Disabled test spec/ruby_llm/protocols/vertexai/ranking_spec.rb:41
- Disabled test spec/ruby_llm/providers/azure/cohere_spec.rb:179
- Disabled test spec/ruby_llm/providers/bedrock/mantle_spec.rb:211
- Disabled test spec/ruby_llm/providers/elevenlabs_spec.rb:101
- Disabled test spec/tasks/mutation_testing_spec.rb:155
- Silent fallback default on parse failure lib/ruby_llm/code_execution/packages.rb:65
- Silent fallback default on parse failure lib/ruby_llm/mcp/content.rb:44
- Silent fallback default on parse failure lib/ruby_llm/protocols/converse/streaming.rb:455
- Silent fallback default on parse failure lib/ruby_llm/tools/search.rb:25
- Projects may be oversized for their cohesion
- REDACTED
- REDACTED
- REDACTED
- No ADRs
- Ruby's type checker does not check nil everywhere
Appendix B — Reproduction & audit trail
| Dimension | Tool | Version | Command | Findings | Raw output |
|---|---|---|---|---|---|
| D28 · Secrets (history) | gitleaks | — | gitleaks detect --no-banner --config /opt/gitleaks-rules/watchdog-gitleaks.toml --report-format json --report-path /tmp/watchdog-gitleaks-e5e6c7fc2a96429daf111bddfc7126f4/history.json --exit-code 0 --source . | 0 | — |
| D29 · Static Analysis (SAST) | semgrep | — | semgrep --config /opt/semgrep-rules/security-audit.yml --config /opt/semgrep-rules/owasp-top-ten.yml --config /opt/semgrep-rules/watchdog-sast.yml --json --quiet --timeout 10 --timeout-threshold 3 --metrics off --exclude *.arrow --exclude *.avro --exclude *.csv --exclude *.dat --exclude *.db --exclude *.feather --exclude *.geojson --exclude *.h5 --exclude *.hdf5 --exclude *.jsonl --exclude *.log --exclude *.ndjson --exclude *.npy --exclude *.npz --exclude *.orc --exclude *.parquet --exclude *.pickle --exclude *.pkl --exclude *.psv --exclude *.sqlite --exclude *.sqlite3 --exclude *.tfrecord --exclude *.tsv --exclude *.txt --exclude media . | 60 | artifacts/raw/semgrep.json |
| D30 · Dependency Vulnerabilities | osv-scanner | — | osv-scanner --format json --recursive --experimental-disable-plugins transitivedependency/requirements --experimental-disable-plugins dotnet/csproj --experimental-disable-plugins dotnet/nugetcpm . | 0 | — |
| D30 · Dependency Vulnerabilities | trivy | — | trivy fs --scanners vuln --format json --quiet --severity CRITICAL,HIGH,MEDIUM,LOW --skip-dirs **/bin/** --skip-dirs **/obj/** . --skip-db-update | 0 | artifacts/raw/trivy-fs.json |
| D31 · IaC & Container Security | trivy | — | trivy: not applicable — No Infrastructure-as-Code or container manifests found (Dockerfile, Docker Compose, Terraform, Kubernetes/Helm, CloudFormation, ARM, Bicep, Ansible); nothing to scan. | 0 | — |
| D32 · Data Compliance (PII/GDPR) | semgrep | — | semgrep: not applicable — No personal data was found crossing a boundary the PII/GDPR ruleset checks — nothing written to a log or console sink, placed in a URL or query string, or persisted to browser storage. That is a clean result for the LEAK surface only: this ruleset detects personal data escaping, it does not inventory the personal data a repository holds, so it is not evidence that this repository has no personal-data surface. The personal-data map (Appendix C) and the C1-C5 compliance cards are what speak to that. semgrep could not parse 1 file(s) — `lib/ruby_llm/provider.rb` — so the PII/GDPR sweep did not cover the unparsed regions of them; rows reported elsewhere in those files are real. | 0 | — |
| D37 · Vulnerability-disclosure Policy | disclosure | — | disclosure: not applicable — No vulnerability-disclosure policy file found (SECURITY.md/.markdown/.rst/.txt at root or under .github/.forgejo/.gitea/docs, .well-known/security.txt). A coordinated-disclosure policy may live off-repo, so this is not evidenced rather than failed. | 0 | — |
| D40 · Network Egress Confinement | runtime-hardening | — | runtime-hardening: not applicable — No Kubernetes/orchestration workloads found in the repository manifests; network egress policy is a cluster-native control that may live at the platform/firewall layer, so there is nothing to assess here. | 0 | — |
| D41 · Kernel & Syscall Confinement | runtime-hardening | — | runtime-hardening: not applicable — No Kubernetes/orchestration workloads found in the repository manifests; seccomp/AppArmor/SELinux confinement is a workload-level control, so there is nothing to assess here. | 0 | — |
| D42 · Runtime Threat Enforcement | runtime-hardening | — | runtime-hardening: not applicable — No Kubernetes/orchestration workloads found in the repository manifests; runtime threat-detection and admission-control policy are cluster-level controls, so there is nothing to assess here. | 0 | — |
| D43 · Malicious Dependencies | osv-scanner | — | osv-scanner --format json --recursive --experimental-disable-plugins transitivedependency/requirements --experimental-disable-plugins dotnet/csproj --experimental-disable-plugins dotnet/nugetcpm . | 0 | — |
| D43 · Malicious Dependencies | trivy | — | trivy fs --scanners vuln --format json --quiet --severity CRITICAL,HIGH,MEDIUM,LOW --skip-dirs **/bin/** --skip-dirs **/obj/** . --skip-db-update | 0 | artifacts/raw/trivy-fs.json |