Problem
Snouty currently fails when agents request paths like generated/sdk/go/assert because these docs are not produced by Eleventy and therefore don’t exist in the indexed markdown set.
Context
Agents frequently attempt to access generated SDK documentation. Since this content is not available in Snouty’s current pipeline, it results in errors or missing data during retrieval.
Proposed Solutions
- Option A: Ingest generated SDK docs into Snouty’s dataset
Generate markdown versions of SDK docs (e.g., from the existing generator pipeline)
Include them in the indexing process alongside Eleventy output
- Option B: Add redirect or fallback logic in Snouty
Detect requests for generated paths (e.g., generated/sdk/...)
Redirect agents to an alternative source or structured representation
Challenges
Generated SDK docs are not currently available in markdown format
Requires integrating a new content source into the pipeline or modifying Snouty’s resolution logic
Open Questions
What is the best way to produce markdown versions of generated SDK docs?
Should Snouty treat generated content as first-class indexed data or handle it via redirects?
Problem
Snouty currently fails when agents request paths like generated/sdk/go/assert because these docs are not produced by Eleventy and therefore don’t exist in the indexed markdown set.
Context
Agents frequently attempt to access generated SDK documentation. Since this content is not available in Snouty’s current pipeline, it results in errors or missing data during retrieval.
Proposed Solutions
Generate markdown versions of SDK docs (e.g., from the existing generator pipeline)
Include them in the indexing process alongside Eleventy output
Detect requests for generated paths (e.g., generated/sdk/...)
Redirect agents to an alternative source or structured representation
Challenges
Generated SDK docs are not currently available in markdown format
Requires integrating a new content source into the pipeline or modifying Snouty’s resolution logic
Open Questions
What is the best way to produce markdown versions of generated SDK docs?
Should Snouty treat generated content as first-class indexed data or handle it via redirects?