agenttwin-prepare
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external application code and imported helpers, creating an ingestion surface for untrusted data that could potentially influence agent behavior.\n- Ingestion points: Application entry points and source code analyzed during discovery in SKILL.md.\n- Boundary markers: No explicit delimiters are defined to separate untrusted codebase content from the agent's control logic.\n- Capability inventory: The instructions imply the agent will perform filesystem reads/writes, network operations, and command execution.\n- Sanitization: The skill relies on manual inspection ('inspect these before uploading') but provides no automated sanitization for the ingested code content.
Audit Metadata