skills/outsharp/shipp-skills/shipp/Gen Agent Trust Hub

shipp

Pass

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill processes real-time event data from external API endpoints, which presents a surface for indirect prompt injection. Content within the external data stream could potentially be crafted to influence the agent's behavior.\n
  • Ingestion points: The skill retrieves dynamic content via https://api.shipp.ai/api/v1/connections/{connectionId} and https://api.shipp.ai/api/v1/sports/{sport}/schedule.\n
  • Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are implemented in the skill's data handling definitions.\n
  • Capability inventory: The skill utilizes curl for network communication and jq for processing JSON payloads.\n
  • Sanitization: The skill does not define specific sanitization or isolation logic to prevent the model from interpreting retrieved data as instructions.\n- [EXTERNAL_DOWNLOADS]: The skill is configured to fetch data and documentation from the vendor's official API and documentation domains.\n
  • Evidence: The allowed-tools section grants access to https://api.shipp.ai/*, https://docs.shipp.ai/*, and https://platform.shipp.ai/* using curl.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 8, 2026, 06:56 PM