pw-coverage-planner

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from the user and the local codebase.
  • Ingestion points: The skill reads user-provided scenarios from the $ARGUMENTS variable, existing test files in the tests/ directory (including sensitive files like auth.setup.ts), and application source code across various languages and frameworks to identify testing requirements.
  • Boundary markers: The instructions do not specify the use of delimiters or provide explicit warnings to the agent to disregard instructions potentially embedded in the source code or existing tests while performing its analysis.
  • Capability inventory: The skill can read file contents, write markdown artifacts to the .claude/pw-plans/ directory, and interact with the user via the AskUserQuestion tool. It utilizes a restricted 'plan mode' (read-only) for research activities.
  • Sanitization: No sanitization, escaping, or validation is applied to the content read from the filesystem before it is interpolated into the reasoning process used to generate the test coverage plan.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 11:30 PM
Security Audit — agent-trust-hub — pw-coverage-planner