The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the agent to execute various shell commands to inspect the repository state and history.
Evidence: In SKILL.md, the instructions mandate running git status, git diff --cached, git log, and git config user.name to gather necessary context.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the repository's staged changes, which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The agent reads content from git diff --cached and reads repository files to clarify intent as described in SKILL.md.
Boundary markers: Absent. There are no instructions provided to the agent to distinguish between code content and potential embedded instructions within the diffs.
Capability inventory: The skill has access to shell execution (git commands) and file system read access.
Sanitization: Absent. The skill does not specify any validation or sanitization for the content retrieved from the Git diffs or files.

conventional-commit