FlowCouncil — AI Think Tank With 5 Expert Fellows
PassAudited by ClawScan on May 1, 2026.
Overview
This is a coherent instruction-only AI debate skill with no code, credentials, persistence, or destructive behavior; users should mainly notice that Deep mode may perform web searches and the expert panel is simulated.
This skill appears safe to install as an instruction-only debate and decision-support prompt. Before using it, remember that Deep mode may browse the web, and the named Fellows are simulated expert personas rather than verified professionals.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If you choose Deep mode, the agent may browse the web and incorporate external sources into the discussion.
Deep mode asks the agent to use web search tools. This is disclosed, scoped to research questions, and aligned with the skill's research-backed debate purpose.
Run web searches for each question. Prioritize: Specific data points with dates and sources
Use Quick mode if you do not want web research, and review cited sources before relying on the final verdict.
The advice may sound like it comes from real domain experts, even though the Fellows are generated personas.
The skill assigns detailed expert-like personas to simulated Fellows. The overall artifact frames this as an AI think tank, but the specificity of credentials could make outputs feel more authoritative than they are.
Partner at early-stage fund, 12 investments, 2 unicorns. Name: Raj.
Treat the Fellows as structured perspectives, not verified professional advisors, especially for legal, financial, medical, or high-stakes decisions.