swarm-patterns
Pass
Audited by Gen Agent Trust Hub on May 25, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill describes orchestration patterns (e.g., Pattern 4) that interpolate untrusted content directly into sub-agent prompts, creating a vulnerability surface for Indirect Prompt Injection.
- Ingestion points: Research results and file content are interpolated into agent prompts (e.g.,
${research.content}in Pattern 4, and file lists in Pattern 3). - Boundary markers: The provided code snippets do not utilize XML tags, delimiters, or explicit instructions to the sub-agents to ignore potentially malicious content within the interpolated data.
- Capability inventory: The agents used in these patterns include
general-purposeand specialized reviewers which possess capabilities to write tasks, send messages, and potentially execute code or modify the file system depending on their environment. - Sanitization: There is no evidence of data sanitization or validation before interpolation into sub-agent prompts.
Audit Metadata