acp-capability-negotiation
Pass
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains architectural guidance and implementation steps for a commerce protocol. No malicious code, credential harvesting, or unauthorized access patterns were detected.- [EXTERNAL_DOWNLOADS]: The skill uses WebSearch and WebFetch tools to retrieve official RFCs and JSON schemas from well-known and trusted domains (github.com and developers.openai.com). These operations are consistent with the skill's primary purpose of implementing a standard protocol.- [PROMPT_INJECTION]: The skill was analyzed for indirect prompt injection surfaces. While it fetches data from external URLs, the instructions target reputable documentation sites, and no attempt to override agent safety filters or system instructions was found.
Audit Metadata