a2a-agent-card
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch technical specifications and real-world examples from external, third-party sources (a2a-protocol.org and GitHub) to inform its output. This creates a surface for indirect prompt injection where untrusted content from those sites could attempt to influence the agent's behavior.
- Ingestion points:
SKILL.md(via instructions to useWebFetchforhttps://a2a-protocol.org/latest/specification/andWebSearchfora2aprojectschemas and samples on GitHub). - Boundary markers: Absent; the instructions do not include specific delimiters or warnings to ignore potential instructions embedded within the fetched documentation.
- Capability inventory:
Bash,Write,Edit,Read,Glob,Grep,WebSearch,WebFetch. - Sanitization: Absent; the agent is expected to parse and use the external schema and documentation directly to generate JSON documents.
Audit Metadata