claude-code-agent-teams

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [Safe] (SAFE): No malicious behaviors, obfuscation, or unauthorized data access patterns were identified. The content is purely instructional.
  • [No Code] (SAFE): The skill consists exclusively of Markdown documentation files and does not include any executable scripts, binaries, or active code components.
  • [Indirect Prompt Injection] (SAFE): The described workflows involving multi-agent coordination for tasks like code review and research create a surface for indirect prompt injection via ingested external data. This is documented as a functional aspect of the feature rather than a specific vulnerability in the skill itself.
  • Ingestion points: PR diffs, research documents, and codebase files processed by teammates.
  • Boundary markers: No explicit delimiters are suggested in the example prompts, relying on the tool's natural language processing.
  • Capability inventory: Teammates possess standard tool access (bash, file editing) as part of the Claude Code environment.
  • Sanitization: The skill relies on the underlying LLM's core safety and filtering protocols.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:13 PM