ramp-install-auth

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill interacts with the Ramp API and processes its JSON responses, creating a surface where untrusted data enters the agent context.
  • Ingestion points: Data from token_resp.json() and resp.json() in SKILL.md.
  • Boundary markers: None provided to delimit the data processed from the API.
  • Capability inventory: Network access via the requests library in SKILL.md.
  • Sanitization: No sanitization or validation of API response content is demonstrated.
  • [EXTERNAL_DOWNLOADS]: The skill makes network requests to obtain authentication tokens and fetch card data.
  • Fetches information from sandbox-api.ramp.com and api.ramp.com, which are the official endpoints for the well-known Ramp service.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 10:51 PM
Security Audit — agent-trust-hub — ramp-install-auth