The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell utilities including rg, test, and readlink to verify repository structure and the validity of paths mentioned in documentation (SKILL.md).
[PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by processing repository documentation files which are untrusted external inputs.
Ingestion points: Reads AGENTS.md, README.md, CONTRIBUTING.md, SECURITY.md, and content within docs/ (SKILL.md).
Boundary markers: Absent; the skill does not instruct the agent to ignore or delimit embedded instructions within the documents being audited.
Capability inventory: The skill allows reading and overwriting documentation files throughout the repository (SKILL.md).
Sanitization: Absent; no content validation or filtering is specified for the audited documentation.
[EXTERNAL_DOWNLOADS]: The reference documentation links to resources from reputable organizations including OpenAI, Stripe, and Anthropic (references/documentation.md).

docs