skill-manager

Pass

Audited by Gen Agent Trust Hub on Apr 15, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: The skill provides automated management for the agent's skill ecosystem, including gap analysis, template-based generation, and quality validation.
  • [EXTERNAL_DOWNLOADS]: The skill references and reads documentation from the vercel-labs-agent-skills repository on GitHub to provide analysis context and templates.
  • [PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection in its skill generation mode. 1. Ingestion points: User-provided free-form descriptions in MODE 2. 2. Boundary markers: None explicitly defined for input interpolation. 3. Capability inventory: File-write operations to .skills/custom/ and .skills/AGENTS.md. 4. Sanitization: Health Check (MODE 4) validates the generated Markdown for structure, length, and locale (en-AU).
  • [COMMAND_EXECUTION]: All operations are limited to reading and writing Markdown files within the local project structure. No shell commands or subprocesses are executed.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 15, 2026, 08:38 AM