The Agent Skills Directory

[SAFE]: The skill operates entirely within the context of analyzing repository documentation and codebase signals. It does not perform network requests, execute shell commands, or access sensitive system credentials. The instructions focus on terminology alignment and architectural consistency.
[PROMPT_INJECTION]: The skill ingests untrusted data from repository files such as ADRs and markdown documentation. While this technically constitutes an indirect prompt injection surface, the risk is mitigated by the skill's lack of dangerous capabilities and its specific purpose as a documentation analysis tool.
Ingestion points: Reads AGENTS.md, docs/ai/*, ADRs, and general repository documentation as specified in the SKILL.md workflow.
Boundary markers: The instructions do not define explicit boundary markers or delimiters for the ingested content.
Capability inventory: The skill is limited to natural language interaction and suggesting documentation updates; it does not contain subprocess calls, network operations, or file-writing logic.
Sanitization: No specific sanitization or filtering of ingested data is mentioned.

doc-grounded-grilling