The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/git_weekly_report.py executes git log and git rev-parse using subprocess.run. The command arguments are passed as a list, which prevents shell injection vulnerabilities.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it processes untrusted data from git commit logs.
Ingestion points: Commit subjects and bodies are read from git repositories via scripts/git_weekly_report.py and provided to the agent for summarization.
Boundary markers: Absent. The instructions do not specify delimiters to separate commit content from the agent's instructions.
Capability inventory: The skill can read local file metadata via git and write structured reports to the file system. It lacks network capabilities.
Sanitization: The script truncates long commit bodies but does not filter or sanitize the text for potentially malicious instructions.

git-weekly-report