qa-execution
Pass
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed for quality assurance and uses structured references and templates to guide the agent's behavior. No patterns of prompt injection or malicious overrides were found.
- [DATA_EXPOSURE]: The skill handles QA artifacts like screenshots and reports within a user-defined or default temporary directory (
/tmp/qa-execution-*). This is standard behavior for testing tools and does not involve accessing sensitive system files or exfiltrating private data. - [COMMAND_EXECUTION]: The skill utilizes the
agent-browsercompanion tool to perform web interactions. All commands (open, snapshot, click, fill) are legitimate actions for a QA agent and are performed within the context of the user-provided test plan. - [INDIRECT_PROMPT_INJECTION]: The skill ingests test artifacts (plans, journeys) from a previous
qa-reportrun. While this represents a data ingestion surface, the risk is low as the content is expected to be structured test data, and the skill includes procedures to verify the environment before execution.
Audit Metadata