creator-deal-decoder
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted text (legal contracts) provided by users.
- Ingestion points: In
SKILL.md, the "Required Inputs" section directs the agent to process "The contract text (paste it)". - Boundary markers: The instructions do not define clear delimiters (e.g., XML tags or triple backticks) or include specific prompts to ignore instructions that might be embedded within the provided contract text.
- Capability inventory: The skill is limited to generating text analysis and negotiation emails; it does not request tool access, network operations, or file system modifications.
- Sanitization: No input validation or sanitization steps are defined to prevent the agent from being influenced by malicious text formatted as legal clauses within the contract input.
Audit Metadata