The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to audit and refresh documentation by systematically reading files across a repository. This creates a surface for indirect prompt injection where malicious instructions embedded in the analyzed repository's files (READMEs, docstrings, or configuration) could influence the agent's behavior.
Ingestion points: Discovery and inventorying of every documentation file in the repo (docs/, root *.md, AGENTS.md, README.md, etc.) as described in CHECKLIST.md Phase 1.
Boundary markers: The instructions do not specify the use of delimiters or provide warnings to the agent regarding untrusted content within the files being audited.
Capability inventory: The agent is granted capabilities to read, write, and delete files, as well as execute shell commands discovered within the repository (git log, lint/test scripts).
Sanitization: There is no mention of sanitizing or validating external content before it is processed or rewritten into new documentation.
[COMMAND_EXECUTION]: The skill instructs the agent to discover and execute project-specific commands for linting, testing, and building (e.g., from package.json, Makefile, justfile, or CI config) as a verification step in SKILL.md and CHECKLIST.md. Executing arbitrary commands found within an untrusted repository poses a risk if the repository contains malicious scripts.

doc-refresh