ielts-reading
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external IELTS reading materials and learner responses, creating an inherent surface for indirect prompt injection. The skill mitigates this by instructing the agent to verify sources and check answers independently.\n
- Ingestion points: IELTS passages and source metadata fetched from external sources.\n
- Boundary markers: None defined in the skill instructions.\n
- Capability inventory: Local filesystem access (read/write/delete) via the included Node.js script.\n
- Sanitization: Structural validation of JSON data is performed, but natural language content is not sanitized for potential embedded instructions.\n- [COMMAND_EXECUTION]: The skill utilizes a bundled Node.js script to manage persistent state. This script is securely designed, using only built-in modules and enforcing a restrictive file creation mask (umask 077) to protect learner data from unauthorized access.
Audit Metadata