The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests review reports and chat-based findings to generate and execute file edit plans, which could allow malicious instructions in the source data to influence agent behavior.
Ingestion points: Data is read from external skill-review reports and inline chat history as specified in SKILL.md.
Boundary markers: The skill employs a structured planning schema (references/plan-format.md) but lacks specific boundary delimiters to isolate the agent from executing instructions potentially embedded within the review text.
Capability inventory: The skill utilizes Write, Edit, and Bash tools, granting it the ability to modify any file within the project scope.
Sanitization: Risk is mitigated by a mandatory process requirement to walk the user through each individual edit for manual confirmation before application.
[COMMAND_EXECUTION]: The skill invokes local shell scripts and third-party development tools to manage skill state and validate changes.
Executes scripts/skill-health.sh to retrieve telemetry data in JSON format for prioritization.
Executes skills/workflows/sync-skills/scripts/sync-skills.sh to distribute validated skills to global locations.
References the use of npx markdownlint-cli2 within its validation checklist (references/validation-checklist.md) to ensure markdown quality.

skill-update