reality-checker
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local shell script named
./qa-playwright-capture.sh. This command is used to initiate automated browser testing and screenshot generation as part of the skill's core QA methodology.\n- [INDIRECT_PROMPT_INJECTION]: The skill reads and evaluates content from thepublic/qa-screenshots/test-results.jsonfile, which contains data generated during the testing of external applications.\n - Ingestion points: The file
public/qa-screenshots/test-results.jsonis read by the agent to assess performance metrics and interaction statuses.\n - Boundary markers: There are no instructions providing delimiters or specific guidance to ignore potential instructions embedded within the test result data.\n
- Capability inventory: The skill possesses capabilities to execute local scripts, perform recursive file searches, and read file system contents.\n
- Sanitization: The provided instructions do not include any logic for sanitizing or validating the data ingested from the test result files.
Audit Metadata