diagnose-issue
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The workflow explicitly authorizes the agent to execute shell or runtime commands to validate issues (e.g., 'Use runtime commands only when they are cheap, local, and useful for diagnosis'). While the instructions specify 'local' use, this capability enables the execution of commands derived from context within the project environment.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external or untrusted sources within the repository, such as source code comments, logs, and documentation, which could contain malicious instructions targeting the AI agent.
- Ingestion points: Source code files, project documentation (ADRs), error logs, and issue descriptions containing Jira links or feature names.
- Boundary markers: The instructions do not specify the use of delimiters or 'ignore' patterns to prevent the agent from following instructions embedded in the data it is analyzing.
- Capability inventory: The skill allows the agent to read arbitrary files in the checkout and execute runtime commands for validation.
- Sanitization: There are no documented steps for sanitizing, escaping, or filtering the content retrieved from the codebase before the agent processes it for the final 'Problem Definition'.
Audit Metadata