swarm-roundtable-orchestrator

Pass

Audited by Gen Agent Trust Hub on Apr 14, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: Analysis of the orchestration scripts and workflow instructions confirms that the skill operates as intended for multi-agent coordination without malicious intent or security vulnerabilities.\n- [PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection through task orchestration.\n
  • Ingestion points: The scripts/swarm_roundtable_plan.py script accepts user input via the --task and --context arguments.\n
  • Boundary markers: Prompts generated in build_spawn_prompt and build_roundtable_prompt do not include explicit 'ignore embedded instructions' markers.\n
  • Capability inventory: The skill executes local Python scripts for planning and consensus calculation, and generates configuration files for sub-agents.\n
  • Sanitization: Task and context strings undergo minimal whitespace sanitization before being interpolated into sub-agent prompts.\n- [COMMAND_EXECUTION]: The scripts/swarm_roundtable_plan.py script dynamically generates agent configuration files (.toml) and workflow documents (.md) based on predefined internal templates to facilitate the multi-agent runtime environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 14, 2026, 09:24 AM
Security Audit — agent-trust-hub — swarm-roundtable-orchestrator