The Agent Skills Directory

[COMMAND_EXECUTION]: The Python scripts in the scripts/ directory, specifically scripts/snapshot.py, scripts/test_snapshot.py, and scripts/trial_fixture.py, use the subprocess module to interact with the local git binary. These executions are used to retrieve commit metadata and generate repository diffs. The implementation uses list-based arguments and incorporates path validation to mitigate command injection risks.
[DYNAMIC_EXECUTION]: The skill utilizes runpy.run_path within its test and fixture scripts to load and execute functions from scripts/snapshot.py. This dynamic execution is confined to the skill's own directory for local code reuse.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted code delivered for review, creating a surface for indirect prompt injection. 1. Ingestion points: File contents are read into fingerprints by scripts/snapshot.py and inspected by the agent during review. 2. Boundary markers: The SKILL.md instructions mandate separate, independent read-only contexts for Standards and Spec axes. 3. Capability inventory: The skill can execute git commands and perform file writes within its own test infrastructure. 4. Sanitization: The snapshot script performs canonical path validation to prevent directory traversal attacks during file inspection.

review-delivery