playwright-validation

Pass

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: SAFECREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill contains hardcoded default credentials (admin@open-metadata.org / admin) for the OpenMetadata application. While these appear to be standard defaults for local development and testing environments (e.g., localhost:8585), hardcoding credentials in instructions is a documented security concern.
  • [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection due to its core functionality.
  • Ingestion points: The skill reads local source code files (.tsx, .ts, .css) via git diff and component file reads. It also ingests web page content during the validation phase using mcp__playwright__browser_snapshot.
  • Boundary markers: The instructions do not define clear boundaries or provide explicit "ignore" directives to prevent the agent from obeying instructions that might be embedded in the code or the web application being tested.
  • Capability inventory: The agent is authorized to execute shell commands (git, yarn) and is instructed to write and modify local test files.
  • Sanitization: There is no evidence of sanitization or filtering of the content ingested from the codebase or the local web server before it is processed by the agent.
  • [COMMAND_EXECUTION]: The skill uses shell commands to interact with the local repository, including git diff for change analysis and yarn lint:playwright for code quality checks. These are standard development operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 7, 2026, 03:30 PM