The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the local working tree through git operations, creating a surface for indirect prompt injection if malicious instructions are embedded in the code being reviewed.
Ingestion points: The skill ingests data from the working tree via git status, git diff, and git diff --cached as described in SKILL.md (Steps 1 and 4).
Boundary markers: Absent. The instructions do not define delimiters or warnings to isolate the code content from the agent's internal logic.
Capability inventory: The skill utilizes shell access to execute git commands and project-specific verification tools (unit tests, lint, build) as noted in SKILL.md (Steps 1, 3, 4, 6, and 7).
Sanitization: Absent. The workflow does not include steps to sanitize or filter the content of the diffs before the agent interprets them for commit message generation.
[COMMAND_EXECUTION]: The skill's primary function involves executing a series of shell commands to manage the git workflow and verify code quality.
Evidence: SKILL.md explicitly directs the agent to run git status, git diff, git add, git commit, and arbitrary project verification commands such as linting, testing, or building in Step 7.

commit-work