mage-hand
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to manipulate external data sources, creating an attack surface for indirect prompt injection if those sources contain malicious instructions.
- Ingestion points: File paths, ticket IDs, record keys, and field names (SKILL.md).
- Boundary markers: The procedure requires the agent to restate targets, success conditions, and no-touch boundaries before taking action to ensure scoping.
- Capability inventory: The skill is scoped for surgical edits to files and record states, implying file-write and data modification capabilities.
- Sanitization: The skill does not define technical sanitization; instead, it relies on procedural guardrails such as diff proofs and reporting exceptions to maintain integrity.
Audit Metadata