The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze user-provided documentation drafts. This creates a potential surface for indirect prompt injection if a draft contains malicious instructions. However, the risk is minimal as the skill's capabilities are limited to reading local reference files and appending to a local review log.
Ingestion points: Processes external documentation drafts provided in the agent's context.
Boundary markers: None explicitly defined in the instructions to separate user data from agent instructions.
Capability inventory: File read access for reference guides (voice-guide.md, terminology.md, mdx-authoring.md, diagrams.md) and file write/append access for the review log (.agents/review-log.md).
Sanitization: No explicit sanitization or filtering of the ingested draft content is mentioned.
[SAFE]: The use of local file references for style guides and terminology is a standard practice for specialized AI agent skills and does not involve external data exfiltration.
[SAFE]: The 'Review Log' feature, which appends metadata to .agents/review-log.md, is a benign persistence of activity for auditing purposes and does not constitute a malicious persistence mechanism.

docs-reviewer