The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes git commands and a local Python utility script (scripts/list_ranges.py) using the Bash tool. The Python script invokes subprocess.run using a list of arguments without enabling shell=True, which is a secure implementation that prevents shell injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill processes git commit history, which can contain untrusted data from any contributor to the repository.
Ingestion points: Commit messages are retrieved via git log and passed directly into the prompts for subagents in SKILL.md (Step 4).
Boundary markers: While the subagent prompt provides specific instructions for formatting the markdown block, it lacks explicit delimiters or instructions to ignore potential commands embedded within the commit subjects themselves.
Capability inventory: The skill possesses the Write and Edit capabilities to modify the project's filesystem and the Bash capability to execute shell commands.
Sanitization: No sanitization, escaping, or filtering of commit message content is performed before the data is interpolated into the agent's context.

make-changelog