diagnosing-bugs
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to generate and execute various commands to build feedback loops, including CLI invocations, HTTP scripts, and 'throwaway harnesses'. This is an intended capability for a software development and debugging tool.
- [PROMPT_INJECTION]: The skill processes external data such as network traces (HAR files), event logs, and core dumps, creating a surface for potential indirect prompt injection attacks.
- Ingestion points: Artifacts such as HAR files, log dumps, core dumps, and error messages (SKILL.md, Phase 1 and Phase 2).
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified when processing these artifacts.
- Capability inventory: The agent has the capability to write files, execute shell commands (bash, curl, tests), and perform network operations as part of the diagnosis loop.
- Sanitization: No explicit sanitization or filtering of external content is mentioned before processing.
Audit Metadata