qa-playswag
Fail
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches OpenAPI/Swagger specifications from arbitrary remote URLs provided by the user using
curlorwgetutilities across the JavaScript, Python, and TypeScript analyzer scripts.\n- [COMMAND_EXECUTION]: The analyzer scripts rely onsubprocess.run(Python) andexecSync/execFileSync(Node.js) to execute shell commands for fetching remote specs and performing format conversions.\n- [REMOTE_CODE_EXECUTION]: Scanners detected a pattern where remote data is downloaded and immediately processed via subprocess-invoked interpreters (node -eorpython3 -c) to handle YAML parsing. While intended for data processing, this sequence is a known vector for remote code execution if the input data or parsing environment is exploited.\n- [INDIRECT_PROMPT_INJECTION]:\n - Ingestion points: Ingests external OpenAPI/Swagger specification files in
scripts/analyze.js,scripts/analyze.py, andscripts/analyze.ts.\n - Boundary markers: Absent. The skill interpolates extracted spec content directly into output templates.\n
- Capability inventory: Includes file system writing and shell command execution via the
Bashtool.\n - Sanitization: None. Fields such as
operationId,summary, anddescriptionare used verbatim from the spec file in the generatedtasks.mdandreport.htmlfiles.
Recommendations
- HIGH: Downloads and executes remote code from: unknown (check file) - DO NOT USE without thorough review
Audit Metadata