The Agent Skills Directory

[SAFE] (SAFE): No malicious patterns detected. The skill uses standard Git commands and a local Python script for text processing. All file operations are confined to the repository and temporary directory.
[Indirect Prompt Injection] (SAFE): While the skill processes untrusted repository data (git diffs), it does so using a dedicated parsing script that validates hunk indices and avoids direct shell execution of the data. 1. Ingestion points: extract-hunks.py reads patch files generated from the repository. 2. Boundary markers: The script relies on the structured nature of git patches (@@ headers). 3. Capability inventory: Uses git apply --cached and git commit on selected hunks. 4. Sanitization: The Python script validates hunk selections and only outputs formatted patch segments, mitigating risks associated with malicious diff content.

commit-splitter