workflow-feature-docs

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill uses emphatic and non-standard phrasing like "IMPORTANT MUST ATTENTION" and "[BLOCKING]" to enforce strict adherence to workflow steps. These instructions are designed to prevent the agent from skipping validation gates rather than bypassing core safety filters.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by design, as it processes user prompts as context for a workflow that has file manipulation and task management capabilities.
  • Ingestion points: The user prompt is ingested via the /workflow-start command to provide context for the documentation process in SKILL.md.
  • Boundary markers: No specific delimiters or "ignore previous instructions" guards are implemented to separate user content from the workflow's template requirements.
  • Capability inventory: The workflow enables the agent to perform file system analysis (/scout, /investigate), generate tasks (TaskCreate), and modify documentation files (/docs-update).
  • Sanitization: No input validation or content filtering mechanisms are defined to sanitize user-provided context before it influences tool usage.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 09:18 AM
Security Audit — agent-trust-hub — workflow-feature-docs