pr-message-writer
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to interact with the local git repository for context gathering.
- Evidence: Use of
git symbolic-ref,git merge-base,git log, andgit diffinSKILL.mdto identify branch changes. - Purpose: This is the primary function of the skill to analyze code differences for PR documentation.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from code changes and external tickets which could theoretically contain malicious instructions.
- Ingestion points: Code diffs (retrieved via
git diff), git logs, and user-provided ticket content. - Boundary markers: The skill does not explicitly define delimiters for this data, but provides clear instructions to the agent on how to interpret it.
- Capability inventory: The skill has the ability to execute git commands and generate text output.
- Sanitization: The skill includes a 'Quality Checklist' that explicitly instructs the agent to ensure no sensitive data (secrets, tokens) is included in the final PR message.
- [SAFE]: The skill references an external repository owned by the author.
- Evidence: Link to
https://github.com/krzysztofsurdy/code-virtuosoinreferences/template.md.
Audit Metadata