agenttwin-investigate
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where the agent processes external, untrusted data including application receipts, workflow definitions, and execution logs to diagnose and repair issues.
- Ingestion points: External receipts, workflows, and application output logs described in
SKILL.md. - Boundary markers: The instructions do not specify any delimiters or safety markers to isolate potentially malicious data in the input.
- Capability inventory: The agent is authorized to modify local application code (patching) and execute shell commands via the
agenttwinCLI tool. - Sanitization: The instructions explicitly direct the agent to remove secrets and private records from reports before sharing, providing a measure of data protection.
Audit Metadata