The Agent Skills Directory

[PERSISTENCE_MECHANISMS]: The skill instructs the installation of a git pre-commit hook by copying a script from the hooks/ directory to .git/hooks/pre-commit. This creates a persistence mechanism where the script executes automatically during the local git workflow.
[PRIVILEGE_ESCALATION]: The skill uses chmod +x to modify permissions for the git hook script. While standard for hook installation, this pattern of granting execution permissions to internal scripts is noted.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it is designed to ingest and process untrusted data from project files and git diffs.
Ingestion points: Git staged diffs and user-specified files (e.g., app.py, config.txt).
Boundary markers: None specified in the instructions for handling the content of scanned files.
Capability inventory: Local command execution (python3), file system modification (cp), and permission changes (chmod).
Sanitization: The skill claims to perform redaction of secrets, but the logic resides in the non-provided scripts/guard.py script.

secret-guard