mckinsey-mochary
CLAUDE.md4 starsChanged 7 months agoMITView on GitHub
If you already have a CLAUDE.md, paste this as a section near the top. If you don't, create CLAUDE.md in your project root and paste this in. --> Three skills form the operating foundation for every agent, every skill, and every main-thread task. They are always on. They are not optional. They are the work itself. Together they cover: analytical rigor (problem-solving), execution discipline (work-plan), and operating excellence (mochary-method). Any non-trivial output carries evidence of all three. 7-step Bulletproof Problem…
In this repository3 other files it gives its agents.
- mochary-methodSkill
- problem-solvingSkill
- work-planSkill
The file
We split this CLAUDE.md into the rules it gives an agent
5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 5
If you already have a CLAUDE.md, paste this as a section near the top.
- Distinctive: line 6
If you don't, create CLAUDE.md in your project root and paste this in.
- Distinctive: line 50
Pre-launch checklist: before locking
- Distinctive: line 69
If any answer is no, rewrite.
- Distinctive: line 88
Every strategic output: event plans, weekly priorities, proposals, campaign strategy, monthly reviews, content batches, outreach sequences, …
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<!--
THIS IS A CLAUDE.md SNIPPET. Paste everything below the opening comment
into your project's CLAUDE.md file (at the top is best).
If you already have a CLAUDE.md, paste this as a section near the top.
If you don't, create CLAUDE.md in your project root and paste this in.
-->🧭 HOW WE DO WORK (foundational triad — read first, every time)
Three skills form the operating foundation for every agent, every skill, and every main-thread task. They are always on. They are not optional. They are the work itself.
The Triad
| Skill | Role | When |
/problem-solving | How we THINK | Every complex question, analysis, strategic diagnosis, or "should we do X?" decision |
/work-plan | How we PLAN | Every scoping, projection, sprint plan, or "what are we actually doing?" output |
/mochary-method | How we OPERATE | Every team interaction, decision moment, priority call, meeting, feedback, and emotional regulation check |
Together they cover: analytical rigor (problem-solving), execution discipline (work-plan), and operating excellence (mochary-method). Any non-trivial output carries evidence of all three.
/problem-solving — How we THINK
7-step Bulletproof Problem Solving (Conn & McLean):
- Define the problem (SMART+, porpoise to refine)
- Disaggregate with MECE logic trees
- Prioritize (impact × ability-to-move 2x2)
- Build workplan (hands to /work-plan)
- Analyze (heuristics first: Occam, 80:20, Rule of 72, 5 Whys → big guns if needed)
- Synthesize (pyramid, visualize, cross-cutting insights)
- Tell the story (SCR, lead with resolution)
Cleaving frames: ROIC, price/volume, principal/agent, assets/options, supply/demand, regulate/incent, incidence/severity.
Bias fighting: obligation to dissent, multiple trees, "what would you have to believe?", pre-mortem, distributed voting.
/work-plan — How we PLAN
Required elements for every plan, projection, or strategic output:
- Hypothesis — "If we do X, we get Y because Z"
- Knock-out analysis — the one check that kills the project if it fails. Sequence FIRST.
- SCR one-day answer — Situation / Complication / Resolution
- Funnel or projection with explicit conversion math (not just targets)
- 6-column workplan — Issue / Hypothesis / Analysis / Owner / Output / Due (date + time)
- Kill list — what we're explicitly NOT doing
- Risk register — likelihood × impact × mitigation
- Pre-launch checklist before locking
/mochary-method — How we OPERATE
- Top Goal — 2 uninterrupted hours/day on the single highest priority, calendar-blocked
- Zone of Genius — spend time only on what only you can do; delegate/kill the rest
- DRI — every task, deliverable, decision has one Directly Responsible Individual
- RAPID — Recommend, Agree, Perform, Input, Decide
- Issue / Proposed Solution — state facts, story, proposed action with DRI + due. Never bring an issue without a proposed solution.
- Clean Escalation — facts + options + tradeoffs, not advocacy
- Fear & Anger regulation — name it, feel it, let it go. Don't operate from fear/anger state.
- Feedback model — "I noticed X → I felt Y → I'd love Z"
The 3-question test (run before handing over any output)
- Thinking: Is the analytical structure clear? Problem defined, tree MECE, hypotheses testable, synthesis visible? (
/problem-solving) - Planning: Does every activity tie back to a measurable outcome with a named owner and a date + time? (
/work-plan) - Operating: Is there a clear DRI for every action? Are priorities framed as issues with proposed solutions? Is the framing facts-not-advocacy? (
/mochary-method)
If any answer is no, rewrite.
Anti-patterns we do not accept
- ❌ Logistics checklists disguised as plans
- ❌ Goals without hypotheses
- ❌ Tasks without single named owners (DRI)
- ❌ Due dates without times
- ❌ Outputs that aren't dummied
- ❌ Equal weighting — no knock-out identified
- ❌ Missing post-execution workplan
- ❌ "Boil the ocean" before defining the problem
- ❌ Advocacy framed as analysis
- ❌ Meetings without agendas, decisions without RAPID clarity
- ❌ Issue without proposed solution ("just raising it")
- ❌ Generic narrative instead of pyramid-structured argument
When the triad applies
Every strategic output: event plans, weekly priorities, proposals, campaign strategy, monthly reviews, content batches, outreach sequences, roadmap memos, quarterly planning, team retrospectives, 1:1s, decision escalations.
When it does not apply
Pure production tasks with a defined spec (single tweet, single review, single report to a known format), single-step edits, conversational exchanges.
Why this matters
Most work fails at one of three layers: thinking, planning, or operating. Plans fail because they're logistics-led (work-plan gap). Decisions fail because they're advocacy-led (mochary gap). Analyses fail because they're unstructured (problem-solving gap). The triad forces rigor at all three layers before work becomes expensive to fix.
Full methods: .claude/skills/problem-solving/SKILL.md, .claude/skills/work-plan/SKILL.md, .claude/skills/mochary-method/SKILL.md. Invoked via /problem-solving, /work-plan, /mochary-method.
<!--
END OF CLAUDE.md snippet. Continue with your project's own CLAUDE.md content below.
-->Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.