tidas-contract-context

Pass

Audited by Gen Agent Trust Hub on Jun 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the tiangong-lca command-line interface to fetch datasets and contract artifacts. This involves executing shell commands such as tiangong-lca dataset context-pack and tiangong-lca dataset contract get. These are vendor-provided tools used for the skill's primary purpose.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and directs the agent to follow instructions contained within externally fetched artifacts.
  • Ingestion points: The skill fetches ai-context.md, ai-context.json, methodology.yaml, and schema.json via the CLI, which are then used by the agent to guide data authoring or repair.
  • Boundary markers: No delimiters or instructions are provided to the agent to disregard potential instructions embedded within the fetched TIDAS metadata or context packs.
  • Capability inventory: The agent has the capability to execute CLI commands and interact with the local workspace environment.
  • Sanitization: There is no evidence of validation or sanitization of the fetched content before it is processed by the AI agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 27, 2026, 05:19 AM
Security Audit — agent-trust-hub — tidas-contract-context