coral-create-source-spec

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external provider API documentation and local datasets to generate Coral source specifications. This data is then processed by CLI tools like coral source add and coral sql.\n
  • Ingestion points: API documentation and local datasets are read from the environment as described in SKILL.md.\n
  • Boundary markers: No specific delimiters or instructions are provided to the agent to isolate untrusted data from its core instructions.\n
  • Capability inventory: The skill utilizes coral source lint, coral source add, and coral sql to perform system operations.\n
  • Sanitization: The skill employs coral source lint to validate the generated YAML schema before it is used.\n- [DYNAMIC_EXECUTION]: The skill's primary workflow involves generating configuration files (Coral source specs) and subsequently invoking them through the coral CLI. This pattern of generating and then executing structured content represents a dynamic execution risk.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 10:09 AM
Security Audit — agent-trust-hub — coral-create-source-spec