libertee
by worksystems-designv0.7.1
Structured thinking methods as multi-agent sessions — Six Thinking Hats, Adversarial Debate, Disney Method, Pre-Mortem, Polarity Management, TRIZ, W³, Troika Consulting, Wise Crowds
Documentation
# Libertee Structured thinking methods as multi-agent sessions for Claude Code. Facilitation techniques from the real world — Six Thinking Hats, Adversarial Debate, Disney Creative Strategy, Pre-Mortem Analysis, Polarity Management, TRIZ, W³, Troika Consulting, and Wise Crowds — each powered by specialized AI agents that take on distinct roles and build on each other's insights. No code. No build steps. Just Markdown files that orchestrate multi-agent thinking sessions. ## The Methods ### Six Thinking Hats®  Five perspectives, one at a time. Facts first, then feelings, creativity, opportunities, and risks — in that order. The Blue Hat orchestrator chooses the sequence based on your topic (exploratory, reactive, or cautious) and synthesizes at the end. The separation is the point: mixing perspectives produces muddy thinking, separating them produces clarity. ```bash /libertee:six-hats "Should we adopt a microservices architecture?" ``` --- ### Adversarial Debate  A structured 3-round debate where Pro and Contra must never agree. Each round escalates — opening statements, rebuttals, final arguments — then a Judge delivers the verdict. The agents are instructed to maintain their positions and counter every argument. This prevents the common AI pattern of politely converging and produces genuinely useful tension. ```bash /libertee:debate "Remote work is superior to office work" ``` --- ### Disney Creative Strategy  Three rooms, three mindsets. The Dreamer paints a bold vision without limits. The Realist takes that vision and builds a concrete plan — assuming it IS possible. The Critic stress-tests the plan (not the dream). Disney's genius was separating these modes into different rooms. We separate them into different agents. ```bash /libertee:disney "What if we completely rethought our onboarding?" ``` --- ### Pre-Mortem Analysis  "Imagine it failed. Spectacularly. Now tell me why." The Doom Analyst generates vivid failure scenarios, then reality-checks which ones are already showing early signs. The Facilitator turns it into a prevention plan with the uncomfortable truth nobody wants to hear. Research shows prospective hindsight increases risk identification by 30%. ```bash /libertee:pre-mortem "We're launching a new product in Q3" ``` --- ### Polarity Management®  Not every tension is a problem to solve. Some are polarities to manage — where both sides need each other. Two Pole Advocates each map their pole's upsides AND downsides honestly. The Facilitator synthesizes into a Polarity Map with virtuous cycles, vicious cycles, warning signs, and action steps. Nobody wins. That's the point. ```bash /libertee:polarity "Centralization vs Decentralization" ``` --- ### TRIZ (Liberating Structure)  "What must we do to guarantee the worst possible outcome?" The Saboteur generates creative failure strategies with dark humor, then reality-checks which ones are already happening. Reverse brainstorming at its finest — sometimes the fastest way to improve is to stop making things worse. ```bash /libertee:triz "Our sprint delivery reliability" ``` --- ### W³ — What? So What? Now What? (Liberating Structure)  Three phases, strictly separated. First: what actually happened? (facts only, no interpretation). Then: what does it mean? (patterns and implications). Finally: what do we do now? (concrete actions). The same Reflector agent runs all three phases — the discipline is in the separation. Most teams jump straight to "Now What?" and wonder why nothing changes. ```bash /libertee:w3 "Our last product launch" ``` --- ### Troika Consulting (Liberating Structure)  A Client presents a problem, then two Consultants brainstorm openly while the Client listens in silence. Two rounds of back-and-forth between complementary experts who talk to each other, not to you. The power is in the listening: hearing experts discuss YOUR problem without the temptation to defend, explain, or redirect. By default, all three roles are AI agents — use `--join` to take a seat. ```bash /libertee:troika "We keep missing deadlines despite good planning" ``` --- ### Wise Crowds (Liberating Structure)  4-5 diverse stakeholder perspectives, each seeing all previous contributions. The crowd's value isn't in consensus — it's in the spread. Where they converge reveals what's real. Where they diverge reveals what's interesting. What nobody mentions reveals the blind spot. The Facilitator's synthesis maps convergence, divergence, blind spots, and the signal in the noise. ```bash /libertee:wise-crowds "Should we open-source our internal tooling?" ``` --- ## Features ### Logic Modes All methods support alternative evaluation logic: | Mode | What it does | Flag | |------|-------------|------| | **Binary** | Pick a side (default) | — | | **Tetralemma** | The one, the other, both, neither, or wrong question | `--tetralemma` | | **Polarity** | Map the tension, don't resolve it | `--polarity` | ```bash /libertee:debate "Build vs. buy" --tetralemma ``` ### Join Mode Take on a role yourself — bring your real emotions, domain knowledge, or conviction: ```bash /libertee:six-hats "Team restructuring" --join red # Your real feelings /libertee:debate "Switch to Kubernetes" --join pro # You argue, AI counters /libertee:disney "New onboarding" --join dreamer # Your vision /libertee:pre-mortem "Platform migration" --join doom # You know where it hurts /libertee:polarity "Autonomy vs Alignment" --join a # Your bias, made visible /libertee:triz "Sprint reliability" --join saboteur # Your insider sabotage knowledge /libertee:w3 "Last quarter" --join so-what # Your interpretation /libertee:troika "Missed deadlines" --join client # Your real problem, AI consults /libertee:wise-crowds "Open-source strategy" --join CTO # Your seat at the table ``` ### Brief Mode All methods support `--brief` for tighter output — same structure, same perspectives, fewer words. Ideal for mobile or when you need a quick pulse rather than a deep dive. ```bash /libertee:six-hats "Team restructuring" --brief /libertee:troika "Scaling challenges" --brief --join client ``` ### Session Context Libertee picks up what you discussed before the command. Had a 20-minute conversation about technical debt? `/libertee:pre-mortem` will generate failure scenarios grounded in that context. Want a clean slate? `/clear` first. ### Guide Not sure which method to use? ```bash /libertee:guide "I need to make a tough decision about our tech stack" ``` Works in any language — agents automatically respond in yours. ## Installation ### From GitHub (permanent) ```bash # Add the marketplace (once) /plugin marketplace add worksystems-design/libertee # Install /plugin install libertee@worksystems-design-libertee ``` Restart Claude Code after installation (`/exit`, then relaunch). ### Try it locally (session only) ```bash git clone https://github.com/worksystems-design/libertee.git claude --plugin-dir ./libertee ``` ### Uninstall ```bash /plugin uninstall libertee@worksystems-design-libertee ``` ## How It Works Each method follows the same architecture: 1. **You** provide a topic, thesis, or challenge 2. **The orchestrator** detects your language, picks up session context, and spawns agents one by one 3. **Each agent** receives all previous perspectives as accumulated context, then adds their own 4. **The orchestrator** synthesizes everything at the end — never before all perspectives are in Agents are temporary — they exist only during their turn, deliver their perspective, and are done. The sequential accumulation is the method: each perspective gets richer because it sees everything that came before. ## Examples Complete example sessions (all in `--brief` mode) in English and German: - Six Thinking Hats® — [English](examples/en/six-hats.md) · [Deutsch](examples/de/six-hats.md) - Adversarial Debate — [English](examples/en/debate.md) · [Deutsch](examples/de/debate.md) - Disney Creative Strategy — [English](examples/en/disney.md) · [Deutsch](examples/de/disney.md) - Pre-Mortem — [English](examples/en/pre-mortem.md) · [Deutsch](examples/de/pre-mortem.md) - Polarity Management® — [English](examples/en/polarity.md) · [Deutsch](examples/de/polarity.md) - TRIZ — [English](examples/en/triz.md) · [Deutsch](examples/de/triz.md) - W³ — [English](examples/en/w3.md) · [Deutsch](examples/de/w3.md) - Troika Consulting — [English](examples/en/troika.md) · [Deutsch](examples/de/troika.md) - Wise Crowds — [English](examples/en/wise-crowds.md) · [Deutsch](examples/de/wise-crowds.md) ## Contributing See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines on adding new thinking methods. ## License MIT — see [LICENSE](LICENSE) ## Trademark Notice Methods in this plugin are referenced for educational and descriptive purposes. This plugin is not affiliated with, endorsed by, or sponsored by any of the organizations below. - **Six Thinking Hats®** is a registered trademark of Edward de Bono Ltd. - **Polarity Management®** and **Polarity Map®** are registered trademarks of Barry Johnson & Polarity Partnerships, LLC. - **Disney Creative Strategy** was formalized by Robert Dilts (1994), based on Walt Disney's creative process. - **Pre-Mortem Analysis** is a technique developed by Gary Klein. - **Tetralemma** is rooted in Indian logic, formalized for systemic work by Matthias Varga von Kibed and Insa Sparrer. - **Adversarial Debate** draws on multi-agent debate research (MIT CSAIL, Mitsubishi Electric). - **TRIZ**, **W³ (What? So What? Now What?)**, **Troika Consulting**, and **Wise Crowds** are [Liberating Structures](https://www.liberatingstructures.com/) developed by Henri Lipmanowicz and Keith McCandless. ## Author Thomas Krause — [worksystems.design](https://worksystems.design)