team-ideation_skill

This skill coordinates a multi-agent ideation session to convert a seed concept into actionable briefs without generating ideas itself.
  • Python

100

GitHub Stars

1

Bundled Files

3 weeks ago

Catalog Refreshed

1 month ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstart where the catalogue uses aiagentskills.

npx veilstart add skill sammcj/agentic-coding --skill team-ideation

  • SKILL.md75.8 KB

Overview

This skill launches a structured multi-agent ideation session that pairs a Free Thinker and a Grounder, arbitrated by a Team Lead and documented by a Writer. It turns a seed concept into a set of evaluated, curated idea briefs and produces a reproducible session archive. Use it only when you explicitly request the 'ideation' skill. It supports a "continue" mode to resume and extend previous sessions.

How this skill works

I act as the Arbiter (team lead): I read your seed, create a persistent Agent Team, spawn the Free Thinker, Grounder, and Writer (and an Explorer when research is required), and coordinate their peer-to-peer dialogue. The Thinker and Grounder generate and critique ideas; the Arbiter evaluates reports as "Interesting", "Needs more conversation", or "Not interesting" and routes follow-ups. The Writer documents the ideation graph and produces final briefs; when deliverables are complete I trigger a production phase to generate diagrams, slides, a webpage, and an archive.

When to use it

  • You have a seed idea, prompt, or file and want a focused team to explore directions and produce briefs.
  • You need divergence and convergence separated: creative generation plus rigorous grounding and synthesis.
  • You want a reproducible session archive with captured sources and deliverables.
  • You want to resume or extend prior ideation using the "continue" mode.
  • You expect iterative research requests during ideation and need a dedicated Explorer agent.

Best practices

  • Provide a clear seed (inline text or a path); the Arbiter reads and captures every input into session/sources/
  • Decide whether pre-session, parallel, or no research is needed so the Explorer is spawned appropriately
  • Use short, focused follow-up prompts when asked to refine a specific idea rather than broad restarts
  • Prefer "continue" mode when building on an existing session to preserve context and avoid duplication
  • Trust the Arbiter to delegate; avoid generating ideas yourself during a live team session

Example use cases

  • Turn a one-paragraph product concept into 6–10 evaluated idea briefs and a session vision document.
  • Explore multiple takes on a design problem by letting a Free Thinker propose and a Grounder challenge them.
  • Resume a prior ideation folder to deepen promising directions without restarting the process.
  • Run parallel research on competitors or technical constraints while thinkers continue creative exploration.
  • Produce presentation slides, diagrams, and an archival capsule automatically once the Writer finalizes deliverables.

FAQ

The Arbiter assesses the seed and picks pre-session, parallel, or no research. The Explorer is spawned when background facts are required.

How does "continue" mode find the previous session?

If you pass a path that exists it uses it; if you pass a keyword it searches ideation-*<keyword>* directories and asks you to choose if multiple matches appear.

When does production start?

Production starts after the Writer reports "All deliverables complete"; the Arbiter then spawns four production agents and creates dependent tasks.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational