Agent skill
strategy-salon-alignment
A framework for building "Strategy Salons" (or Nerd Clubs) to drive bottom-up alignment and surface emergent ideas in complex organizations. Use this when formal executive reviews are becoming "organizational kayfabe" (performative but inaccurate), when disparate teams are working on overlapping problems without coordination, or when you need a low-stakes environment to test "ugly baby" ideas before they are ready for formal scrutiny.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/strategy-salon-alignment
SKILL.md
Strategy Salons (Nerd Clubs)
Strategy Salons are "idea labs" designed to terraform the culture of an organization from the bottom up. By creating a high-trust, low-stakes environment, you can bypass the coordination costs and political friction of large organizations to find "miracles"—high-leverage ideas that grow like seeds.
Core Principles
- Design for Emergence: Do not try to force a specific outcome or decision. The goal is to create a space where the most interesting ideas naturally attract more "water" (attention and effort).
- The "Yes, And" Constraint: Adopt an improv-based culture. If an idea is "dumb," do not attack it—simply do not engage with it. If an idea is interesting, build on it.
- Illegibility as Protection: Use names like "Navel Gazers" or "Nerd Club." This signals that the group is for enthusiasts and prevents "party poopers" or status-seekers from joining just for visibility.
- The Gardener Role: You are not a manager; you are a gardener. You prune toxic behaviors and water the seeds of greatness in others.
Implementation Workflow
1. Identify the Seed Crystal
Find a small group (3–5 people) who already naturally discuss the problem space over lunch or in side chats. Ensure these people have an "intrinsically motivated" mindset rather than a "careerist" mindset.
2. Set the "Yes, And" Norms
Explicitly define how the group communicates to avoid the "stabbing in the dark" common in large orgs:
- Use "I wonder": Instead of saying "That won't work because X," say "I wonder if we applied [Lens Y] to that, how it might change." This makes the critique about your perspective, not their failure.
- The Silence Rule: If someone shares an idea that doesn't resonate, the group simply moves on. No one is forced to defend a half-formed idea.
3. Dribble in Perspectives
Slowly add 1–2 new members every few weeks. Prioritize "adjacent possible" perspectives—e.g., add a designer to an engineering group, or a salesperson to a product group.
- The Jerk Test: Avoid adding anyone who asks "What is the ROI of this meeting?" One person can destroy the "Yes, And" energy of the entire group.
4. Build Momentum through Capture
- Public/Private Validation: When someone shares a great idea in a DM, tell them: "That is incredible. You should share that in the group."
- Emoji Rewards: Use heavy emoji reactions to signal what is working.
- The Friday Export: Once a week, summarize the most interesting "butterfly" ideas and share them back to the group. This creates "FOMO" for those who weren't active.
5. Transition to Live Experiments
Once the chat has high momentum, host an optional one-hour "Nerd Session."
- Set the Quorum: Ensure at least 3 high-energy people are guaranteed to show up so the session never feels "dead."
- Post-Session Hype: Share 1-2 "mind-blowing" insights from the session back to the main channel immediately after.
Examples
Example 1: Aligning Overlapping Products
- Context: Twelve different teams at a large tech company are building slightly different developer tools that accidentally undermine each other.
- Application: Start a "DevTools Navel Gazing" group. Instead of forcing a merger, encourage members to share "What if our APIs talked to each other?"
- Output: An emergent "North Star" document that all teams actually believe in because they co-authored it in a low-stakes environment.
Example 2: Identifying AI Opportunities
- Context: A product team is stuck using old playbooks for LLMs and failing.
- Application: Create an "LLM Play Club" where PMs and Engineers share weird WebSim links or odd Claude prompts.
- Output: The team discovers a "Magical Duct Tape" use case for a customer problem that would have been rejected in a formal PRD review for being "too squishy."
Common Pitfalls
- Steering Toward an Outcome: If you try to make the group "produce a roadmap," the "Yes, And" energy dies and it becomes another boring meeting. Let the roadmap be a side effect, not the goal.
- The One-Person Monologue: If only one person (the facilitator) is talking, the community is already dead. You must actively "manufacture" risk-taking by prompting others to share.
- Letting the Kayfabe In: If you invite an executive who demands "rigor" or "metrics" too early, the "ugly baby" ideas will be killed before they can grow. Keep the group "kooky" and "unserious" as long as possible.
- Ignoring Disconfirming Evidence: If the group becomes an echo chamber, it becomes dangerous. Use "I wonder" to introduce dissenting views without breaking the safety of the space.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?