The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the git command and internal Python scripts using the subprocess module to analyze repository history. These calls are implemented using list arguments without enabling shell execution, which is the recommended practice to prevent command injection vulnerabilities.
[EXTERNAL_DOWNLOADS]: The skill identifies networkx as a dependency, which is a well-known, trusted Python library for studying graphs and networks.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from git history (such as filenames, author names, and emails) which is an untrusted external source that could potentially be used for indirect prompt injection.
Ingestion points: Git metadata is ingested via git log in scripts/build_ownership_map.py and scripts/community_maintainers.py.
Boundary markers: The output CSV and JSON files do not utilize specific delimiters to isolate untrusted content for the agent.
Capability inventory: The skill possesses the ability to execute shell commands (git) and read/write local files.
Sanitization: The scripts perform standard string processing but do not include specific sanitization logic for prompt injection payloads within the git metadata.

security-ownership-map