skills/vllnt/skills/manage-skills/Gen Agent Trust Hub

manage-skills

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [NO_CODE]: The skill consists entirely of markdown instructions and procedural guidelines. It does not include any Python scripts, JavaScript files, shell scripts, or other executable binaries.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to create, review, and update other portable agent skills. This ingestion of external skill content creates an attack surface where malicious instructions within a candidate skill could influence the agent during the review or validation phases.
  • Ingestion points: The workflow processes 'portable agent skills' and 'candidate' files during assessment, creation, and update tasks.
  • Boundary markers: The instructions mention a 'portability boundary' for scoping but do not specify technical delimiters or 'ignore' instructions for the agent when processing untrusted skill content.
  • Capability inventory: The skill is intended to perform file-system operations including creating, updating, and retiring (deleting/moving) files based on its workflow.
  • Sanitization: No explicit data sanitization or escaping mechanisms are described; the skill relies on a 'Quality validation' protocol involving human or independent agent review to detect issues.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 08:18 PM