flow-debug
Pass
Audited by Gen Agent Trust Hub on Jul 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a process for bug investigation and routing. It does not contain any hardcoded credentials, malicious command execution, or data exfiltration patterns.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data, specifically stack traces and error logs, which could theoretically contain malicious instructions. However, this is a standard requirement for debugging tools.
- Ingestion points: Reading error logs, stack traces, and codebase changes via git diff.
- Boundary markers: No explicit delimiters are used to separate the untrusted data from the agent's instructions.
- Capability inventory: The skill routes to other automation workflows like
/flow-patchand/flow-spec, but does not execute arbitrary shell commands directly. - Sanitization: No validation or sanitization of the error content is performed before processing.
Audit Metadata