workflow-validate

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues detected. The skill follows standard technical validation principles and Agile Spike methodology.
  • [COMMAND_EXECUTION]: The skill includes instructions to perform validation methods such as "Prototype implementation", "UI/UX verification (Playwright MCP)", and "TDD approach". These actions involve creating and running code within the local environment to verify technical assumptions, which is the primary purpose of the skill.
  • [DATA_EXFILTRATION]: The skill ingests data from a local file (docs/work/{name}/analysis.md) to determine the validation approach. While this represents a data ingestion surface, the access is restricted to the local project structure and does not involve sending data to external or untrusted sources.
  • Ingestion points: Reads docs/work/{name}/analysis.md (SKILL.md).
  • Boundary markers: None explicitly defined for file content.
  • Capability inventory: Capability to modify project files, create files in __prototype__/, and execute shell commands/tests (SKILL.md).
  • Sanitization: None specified for the content of the analysis document.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 09:22 AM
Security Audit — agent-trust-hub — workflow-validate