jj
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [Indirect Prompt Injection] (SAFE): The skill instructs the agent to read and process repository content. 1. Ingestion points: File diffs and commit messages via
jj diffandjj log. 2. Boundary markers: Agent Rule 5 inSKILL.mdrequires single-quoting revsets. 3. Capability inventory: VCS mutations (squash, rebase) and potential alias execution viajj util exec. 4. Sanitization: Explicit instructions for shell quoting and avoiding interactive modes. - [Prompt Injection] (SAFE): No instructions attempt to override agent constraints or bypass safety filters.
- [Remote Code Execution] (SAFE): The skill does not download or execute external scripts.
- [Data Exposure & Exfiltration] (SAFE): No sensitive data access or credential leaks were found.
Audit Metadata