blazing-agents

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a documentation-only resource providing architectural guidance and integration patterns for the Blazing Agents platform. It contains no executable scripts, obfuscated code, or malicious instructions.- [EXTERNAL_DOWNLOADS]: The documentation references official vendor tools, specifically the @blazing-agents/cli via NPM. This is a legitimate vendor resource consistent with the skill's purpose and does not represent a security risk.- [INDIRECT_PROMPT_INJECTION]: The skill acknowledges the risks of processing untrusted data in agent environments and provides specific guidance to mitigate these risks by enforcing security boundaries.
  • Ingestion points: Data enters through agent generation methods (chat, completion, object) and CLI interaction tools (ba chat, ba run).
  • Capability inventory: The platform supports Workspace file operations, MCP tool execution, and asynchronous Task management.
  • Boundary markers: The skill defines "trusted backend boundaries" and "Tenant isolation" as primary security controls.
  • Sanitization: The instructions explicitly warn to keep credentials and secrets out of "model-visible input," "Tool arguments," and "Workspaces."- [CREDENTIALS_UNSAFE]: The skill correctly handles credentials by instructing users to inject the BLAZING_AGENTS_API_KEY via environment variables from a secret manager rather than hardcoding them.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 03:27 AM
Security Audit — agent-trust-hub — blazing-agents