indykite-capture-upsert-relationships

Pass

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/capture.sh uses curl and jq to process data. It includes a host-pinning check that restricts network operations to official IndyKite API domains, preventing accidental or malicious redirection of sensitive credentials.
  • [EXTERNAL_DOWNLOADS]: The skill connects to IndyKite regional APIs for data ingestion. These operations target trusted vendor-controlled infrastructure.
  • [CREDENTIALS_UNSAFE]: The skill handles an API key via environment variables. The associated shell script provides a --print mode that redacts the key, ensuring it is not exposed in logs or console output during testing.
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts user-defined entity and relationship information to build API requests.
  • Ingestion points: Data modeled in SKILL.md (Steps 1 & 2) for relationship properties.
  • Boundary markers: Not explicitly defined, relying on structured JSON format.
  • Capability inventory: Use of curl and jq within the provided shell script.
  • Sanitization: The script uses jq to ensure the generated payload is a valid JSON object before it is sent to the API.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 24, 2026, 07:09 AM
Security Audit — agent-trust-hub — indykite-capture-upsert-relationships