The Agent Skills Directory

[SAFE]: No malicious patterns detected. The skill includes a pre-commit safety guard (scripts/precommit_guard.py) designed to prevent the accidental inclusion of sensitive files like logs, database files, and binaries in Git commits.
[COMMAND_EXECUTION]: The skill executes Git commands (git add, git diff, git commit) to perform its primary function. These commands are executed locally and are preceded by a safety check of the repository state.
[PROMPT_INJECTION]: Potential surface for indirect prompt injection via repository file names. 1. Ingestion points: scripts/precommit_guard.py reads file paths from git status. 2. Boundary markers: Absent; file names are printed directly in the output. 3. Capability inventory: Git staging and commit operations. 4. Sanitization: Absent; file names are not escaped for instruction-like content.

commit-jira