copying-endpoints-across-projects

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the duplication of existing endpoint configurations, creating a surface where malicious HogQL or variable definitions could be transferred and executed in a different project context.
  • Ingestion points: Endpoint configuration data, including queries and variables, is ingested from the source project using the endpoint-get tool as defined in the workflow (SKILL.md).
  • Boundary markers: The instructions do not define strict boundary markers or delimiters for the ingested query content when it is interpolated into the agent's context, though they advise presenting a summary to the user.
  • Capability inventory: The skill utilizes endpoint-create to write configuration to the target project and endpoint-run to execute the query logic for verification (SKILL.md).
  • Sanitization: The instructions require the agent to copy the source query verbatim, meaning no sanitization or filtering is performed on the HogQL content during the migration process.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 07:58 AM
Security Audit — agent-trust-hub — copying-endpoints-across-projects