headless-ghidra-scope
Pass
Audited by Gen Agent Trust Hub on Apr 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified in the skill instructions or workflow.
- [COMMAND_EXECUTION]: The skill uses
ghidra-agent-clito manage scope files within the local project workspace. These tools are used for their intended purpose of defining binary analysis parameters. - [PROMPT_INJECTION]: An indirect prompt injection surface exists where binary inspection results and user guidance are ingested into
scope.yaml. Evidence: 1. Ingestion points: artifacts/pipeline-state.yaml, artifacts/scope.yaml; 2. Boundary markers: absent; 3. Capability inventory: ghidra-agent-cli; 4. Sanitization: absent. This surface is expected for the skill's documented functionality and does not present a significant risk.
Audit Metadata