federal-register-doc-fetch
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
tiangong-aiCLI tool to perform data discovery and retrieval. These operations are restricted to thefederal-register.documentscapability and follow a structured request/response flow using JSON envelopes. - [INDIRECT_PROMPT_INJECTION]: The skill processes metadata retrieved from FederalRegister.gov, which represents an external data ingestion surface.
- Ingestion points: Data enters the agent context via the output of the
tiangong-ai data runcommand. - Boundary markers: The
SKILL.mdfile contains a dedicated 'Result boundaries' section that instructs the agent on how to handle partial, truncated, or blocked results, and warns against treating metadata as legal proof. - Capability inventory: The skill is limited to shell execution of the
tiangong-aiCLI for data retrieval; no file-write or arbitrary network capabilities are present in the provided files. - Sanitization: While no explicit code-based sanitization is shown, the skill provides clear instructional guardrails for the agent to validate claims against official editions.
Audit Metadata