What problem does it solve? Exported symbols, chart values, and build instructions often ship without documentation, forcing future contributors and AI agents to reverse-engineer intent from code. This Skill fills those gaps with comments that are strictly grounded in what the code demonstrably does, and logs anything it cannot truthfully describe for human review. ## Core Features & Use Cases - Multi-language documentation authoring: Adds godoc comments to Go exports, docstrings to public Python functions, helm-docs-style comments to values.yaml entries, and explanatory comments to Dockerfile instructions, shell functions, and GitHub Actions workflow steps. - Groundedness discipline: Every comment must be derivable from reading the code body; symbols that resist truthful description are skipped and logged to deferred-findings memory instead of receiving speculative filler. - Bounded, reviewable batches: Caps each run at 25 candidates, prioritizes high-visibility surfaces, and produces one commit per category so reviewers can revert any batch independently. - Use Case: After a refactor introduces new exported symbols in an operator package, run this Skill to document the undocumented exports, commit the godoc additions separately from any Helm value comments, and delegate the push to a peer agent. ## Quick Start Ask the agent to document the missing docstrings and code comments in the current checkout, keeping each category in its own commit.
