The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes project-related data that could potentially contain malicious instructions from external sources.\n
Ingestion points: The skill reads git log, verify-report.md, and task documentation (04-dev-tasks*.md) to synthesize information.\n
Boundary markers: There are no explicit delimiters defined to separate the data being analyzed from the agent's internal instructions.\n
Capability inventory: The skill uses Read, Write, Glob, Grep, and Edit tools to generate new documentation in the docs/devdocs/patterns/ directory.\n
Sanitization: No specific filtering or sanitization of the input text is mentioned in the workflow.\n- [SAFE]: The skill includes strong operational constraints and oversight mechanisms to mitigate risks.\n
Human-in-the-loop: The workflow requires explicit user confirmation before any documentation is written or updated.\n
Scope limitation: The skill is strictly prohibited from modifying source code files or existing core documents.\n
Permission restriction: It cannot directly modify the agent's memory configuration (AGENTS.md), instead requiring a user-initiated action.

compound