skills/agent-sh/skillers/recommend/Gen Agent Trust Hub

recommend

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill implements an analytical workflow that reads project configuration and workflow history to provide automation recommendations.- [PROMPT_INJECTION]: The skill identifies and mitigates an indirect prompt injection surface associated with processing user conversation observations.
  • Ingestion points: Data is loaded from knowledge/*.json files in the state directory.
  • Boundary markers: The skill relies on structured JSON validation and specific sanitization rules rather than textual delimiters.
  • Capability inventory: The skill discovery phase uses file system globbing to identify existing plugins and components. It produces JSON-based automation templates for external tools.
  • Sanitization: The skill enforces a security policy that filters shell metacharacters and prohibits raw observation text from appearing in executable command scaffolds.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 05:42 AM
Security Audit — agent-trust-hub — recommend