jj-update-pr
Pass
Audited by Gen Agent Trust Hub on May 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands using jj (Jujutsu) and gh (GitHub CLI) to retrieve version control logs, diffs, and to update GitHub pull requests.
- [PROMPT_INJECTION]: The skill reads the output of jj diff and uses it to generate text, creating a vulnerability surface for indirect prompt injection where instructions embedded in code comments could manipulate the agent.
- Ingestion points: Code diffs retrieved via jj diff -r in Step 5.
- Boundary markers: Absent; there are no specific markers or instructions for the AI to ignore embedded content within the diff.
- Capability inventory: Capability to modify pull requests via gh pr edit and update commit messages via jj describe.
- Sanitization: No evidence of input validation or content filtering before the diff data is processed by the AI.
Audit Metadata