The Agent Skills Directory

[SAFE]: No security issues or malicious patterns were identified across any of the analyzed files.
[COMMAND_EXECUTION]: The skill utilizes local shell scripts (scripts/*.sh) to execute standard Git plumbing and porcelain commands such as git fetch, git log, and git fsck. These operations are well-documented, non-destructive by default, and align with the skill's purpose of repository auditing. A Python test suite (tests/test_git_find_all_checkouts.py) also uses the subprocess module to set up Git fixtures; these calls are implemented using safe practices and are non-malicious.
[EXTERNAL_DOWNLOADS]: Certain scripts perform git fetch operations to ensure remote-tracking references are up to date. This is standard functionality for comparing local work against remote branches and targets the user's own configured remotes.
[PROMPT_INJECTION]: The instructions and reference documents provide technical guidance for Git recovery and safety habits. They do not contain any attempts to override model behavior, safety filters, or system instructions.
[DATA_EXFILTRATION]: No unauthorized data access or network transmission patterns were observed. The skill focus is on local Git forensics and interaction with the user's remote repositories for state synchronization.

git-safety-net