agent-tool-contract-review
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill instructions define a workflow for processing untrusted external data, which creates a surface for indirect prompt injection attacks.
- Ingestion points: SKILL_CARD.md identifies tool schemas, command documentation, action inputs, and error logs as primary inputs.
- Boundary markers: The prompt instructions do not specify the use of delimiters or provide warnings for the agent to ignore instructions contained within the reviewed data.
- Capability inventory: The workflow involves writing and adapting skills, suggesting the agent may use file-system modification capabilities to implement the reviewed contracts.
- Sanitization: No validation or sanitization requirements are specified for the external data being reviewed.
- [NO_CODE]: The skill consists entirely of instructional markdown and configuration templates; it does not include executable scripts, binaries, or automated tool invocations.
Audit Metadata