The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/smart_commit.sh is vulnerable to shell command injection. It accepts a commit message as a command-line argument and interpolates it directly into a shell command.
Evidence: MESSAGE="${1:-chore: update code}" followed by git commit -m "$MESSAGE".
Impact: A malicious user or an indirect prompt injection could provide a commit message containing shell metacharacters (e.g., "; rm -rf / #) which would be executed by the bash shell.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection by processing external data (commit messages) that influence shell command execution without sanitization.
Ingestion points: The first argument to scripts/smart_commit.sh (found in SKILL.md usage examples).
Boundary markers: None present to distinguish the commit message from shell control characters.
Capability inventory: Shell execution via git commands.
Sanitization: None; the script uses standard variable interpolation which does not prevent shell breakout.

git-pushing