The Agent Skills Directory

[COMMAND_EXECUTION]: The skill automates development tasks by executing shell scripts (scripts/run_quality_checks.sh, scripts/verify_clean_tree.sh) and git commands (git commit, git add). These actions are local to the repository and consistent with the skill's stated purpose of managing code quality and version control.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its processing of external project data. Ingestion points: The agent reads CHANGELOG.md via scripts/ensure_changelog_unreleased.py and reviews staged code changes via git diff --staged. Boundary markers: No specific delimiters or instructions are used to separate project data from system instructions. Capability inventory: The skill can execute local scripts and git commands. Sanitization: No sanitization of the ingested file content is performed before the agent processes it to generate commit messages or verify state.\n- [SAFE]: No evidence of hardcoded credentials, remote downloads from untrusted sources, persistence mechanisms, or obfuscated code was found in the provided files. The skill relies on well-known local tools and standard library modules.

commit-work