jujutsu
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process external data sources such as version control logs, diffs, and patch files, which could contain malicious instructions aimed at overriding the agent's behavior.
- Ingestion points:
SKILL.mddescribes workflows that ingest data viajj log,jj diff,jj show, andgit applycommands. - Boundary markers: The instructions do not define explicit delimiters or warnings to separate untrusted commit messages or code contents from the agent's control logic.
- Capability inventory: The skill utilizes shell execution capabilities for version control management (
jjandgit). - Sanitization: There is no evidence of sanitization or validation of the content within processed patches or commit logs before they are entered into the agent's context.
Audit Metadata