The Agent Skills Directory

[COMMAND_EXECUTION]: The provided audit_repo.py script executes local git commands (such as git rev-parse, git ls-files, and git log) to gather metadata about the repository. These calls are performed with controlled arguments and are a core part of the tool's intended auditing functionality.
[INDIRECT_PROMPT_INJECTION]: The skill is intended to process untrusted data by reading and analyzing instruction files (AGENTS.md, CLAUDE.md) within a repository. This represents a potential injection surface where malicious repository content could attempt to influence the agent's behavior or report. The skill instructions mitigate this by directing the agent to verify all findings against build files and repository layouts, and to present evidence before recommendations.
[SAFE]: The auditing script implements safety measures to prevent accidental data exposure. It uses regex patterns to identify potential secret keys (like api_key or token) in configuration files but explicitly avoids printing or reporting the associated values. Furthermore, the skill instructions specifically warn the agent never to inspect credential stores.

agentic-readiness