workflow-greenfield-init

Pass

Audited by Gen Agent Trust Hub on Apr 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious code, external downloads, or data exfiltration patterns were detected in the skill instructions or metadata.
  • [PROMPT_INJECTION]: The skill uses authoritative directives (e.g., "MUST", "Do NOT skip") to ensure the agent follows the specific sequence of the project initialization workflow. These are natural operational instructions for a complex task and do not represent a safety bypass.
  • [PROMPT_INJECTION]: The skill ingests user input (the prompt) as the primary context for the greenfield-init workflow. While this serves as an ingestion point for potential indirect prompt injection, it is the fundamental purpose of the skill and relies on the agent's internal guardrails for security.
  • Ingestion points: User prompt (SKILL.md)
  • Boundary markers: Absent
  • Capability inventory: Triggers a chain of internal commands including /web-research, /deep-research, /scaffold, and /test (SKILL.md)
  • Sanitization: None specified in the skill body
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 10, 2026, 07:25 AM
Security Audit — agent-trust-hub — workflow-greenfield-init