commit-messages
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the local repository and processes it to generate text, creating a surface for indirect instructions to influence the agent's output.
- Ingestion points: The skill executes
git diff --cachedandgit diff(SKILL.md, Step 1) to retrieve the content of staged changes. - Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore or escape prompt-like content found within the code diffs.
- Capability inventory: The skill has the capability to write the resulting commit message to a local file,
./commit_msg.txt(SKILL.md, Step 5). - Sanitization: While the skill implements "slop checks" and "hygiene checks" (SKILL.md, Steps 4, 4a) to filter specific words and characters, it does not include sanitization logic to detect or prevent prompt injection attempts embedded in the source code being committed.
Audit Metadata