bailian-model-recommend

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a command (npx skills add) to install a required dependency (bailian-docs-llm-wiki) if it is not detected. This is a standard operation for managing skill components and dependencies within the execution environment.
  • [EXTERNAL_DOWNLOADS]: Dependency installation fetches content from the author's own repository (modelstudioai/skills). This is a legitimate vendor resource used for the skill's primary functionality.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests and processes model description data from a local catalog provided by a peer skill. However, the data is structured (JSON/JSONL) and originates from the same vendor, making the risk low.
  • Ingestion points: models/models.jsonl and models/groups/<slug>.json (SKILL.md, references/data-source.md)
  • Boundary markers: Not explicitly defined in logic, but data is read as structured JSON/JSONL.
  • Capability inventory: Dependency installation via npx (SKILL.md).
  • Sanitization: Not explicitly mentioned; however, recommendations are grounded in the provided catalog data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 07:42 PM
Security Audit — agent-trust-hub — bailian-model-recommend