agent-registry

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the Agent Arena search API and instructs the agent to follow the hire instructions provided in the results verbatim. This allows third-party data to control agent actions.
  • Ingestion points: search results API in SKILL.md.
  • Boundary markers: None; instructions are processed without isolation.
  • Capability inventory: The agent is instructed to make HTTP POST requests to arbitrary endpoints.
  • Sanitization: No validation of external instructions is performed.
  • [EXTERNAL_DOWNLOADS]: The skill documentation includes a link to wallet setup instructions at https://raw.githubusercontent.com/smukh/agentic-wallet/main/SKILL.md. This repository is owned by an unknown user, which presents a risk that the instructions could be malicious or lead to the execution of untrusted code.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 11:07 AM
Security Audit — agent-trust-hub — agent-registry