The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs repository mining by executing git commands through subprocess.run in scripts/gitlog.py. The commands are constructed using list-based arguments rather than shell strings, which is a secure implementation that prevents arbitrary command injection from user-provided repository paths or flags.
[PROMPT_INJECTION]: The skill contains a vulnerability surface for indirect prompt injection (Category 8) because it ingests untrusted data from git logs and source files and asks the agent to interpret them.
Ingestion points: Commit headers, file paths, and file contents are ingested by scripts/mine.py and scripts/gitlog.py.
Boundary markers: The agent instructions in SKILL.md lack explicit delimiters or warnings to ignore instructions that might be embedded in the codebase being analyzed.
Capability inventory: The agent has the ability to run the provided analysis scripts and write output to hotspots/report.md.
Sanitization: The skill does not sanitize or filter commit messages or source code content before presenting it to the agent, allowing a malicious repository to potentially influence the agent's analysis or subsequent actions.

hotspots