ielts-practice
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a local script
scripts/ielts_buddy_api.pyto facilitate API requests. This is the primary method for providing the skill's capabilities and is executed within the local agent environment without remote code loading.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes data from the IELTS Buddy platform, which represents a potential injection surface.\n - Ingestion points: The agent reads practice activity, exam records, and candidate question lists via
scripts/ielts_buddy_api.pyfrom thework.ieltsbuddy.igopx.cnAPI.\n - Boundary markers: The documentation includes explicit instructions for the agent to separate exam recall content from question candidates and to only use authoritative source fields provided by the server.\n
- Capability inventory: The skill is limited to reading practice session metadata, review results, and prediction records. It does not allow for writing to the exam database or modifying user submissions through the agent.\n
- Sanitization: The instructions specify that only server-returned sequences and JSON-formatted arguments should be used for API calls, reducing the risk of processing malformed or malicious inputs.
Audit Metadata