code-generate-document
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the
jj diff -scommand to identify modified files within a Jujutsu version control repository.\n- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface via its code-reading workflow:\n - Ingestion points: Step 3 of the process involves reading and understanding project source code to generate docstrings and ADRs.\n
- Boundary markers: The instructions lack delimiters or explicit warnings to treat ingested code content as untrusted data.\n
- Capability inventory: The skill is authorized to propose modifications to existing files and create new files in the
doc/adr/directory.\n - Sanitization: There is no mention of sanitizing or escaping the content read from files before it is processed by the agent.
Audit Metadata