bailian-knowledge

Pass

Audited by Gen Agent Trust Hub on Jun 20, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Installation instructions include downloading archived releases and cloning source code from GitHub (github.com/konglong87/rag_skills_ali), which is a well-known service for software distribution. These resources are part of the skill's legitimate installation process.
  • [COMMAND_EXECUTION]: The skill executes local, platform-specific binaries (bailian-query-mac, bailian-query-linux, etc.) included in the skill's directory. The source code for these utilities is provided in the scripts/ folder, allowing for auditability and verification of their behavior.
  • [CREDENTIALS_UNSAFE]: The skill requires manual setup of Alibaba Cloud Access Keys in a local config.json file. This is a standard requirement for accessing the Bailian API. The project correctly includes config.json in its .gitignore to prevent users from accidentally committing their secrets to version control.
  • [PROMPT_INJECTION]: The skill includes comprehensive guardrails against indirect prompt injection and hallucinations. It explicitly instructs the agent to discard low-confidence results (score < 0.4) and strictly forbids the agent from providing any information not found in the retrieved knowledge base nodes, protecting the integrity of the output.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 20, 2026, 10:37 PM
Security Audit — agent-trust-hub — bailian-knowledge