decision-frameworks
Skill0 starsChanged 5 months agoMITView on GitHub
Decision-quality reference for boardroom sessions. Use when chairing board deliberations, writing decision memos, evaluating projects, or whenever a business decision needs structure (reversibility, expected value, base rates, bet sizing, tripwires).
6 other projects use a version of this.
Adopted by
Most recently taken up by
- afeezadebayo/thinking-decisions-skillstheir own version, first seen Nov 15, 2026
- listener-He/java-harness-agenttheir own version, first seen Oct 20, 2026
- kumaran-is/claude-code-onboardingtheir own version, first seen May 25, 2026
- michaelaz774/MBraintheir own version, first seen Mar 2, 2026
- Dragoon0x/product-skillstheir own version, first seen Feb 24, 2026
- wpank/aitheir own version, first seen Feb 10, 2026
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- kumaran-is/claude-code-onboarding★ 35same name
- listener-He/java-harness-agent★ 16same name
- wpank/ai★ 11same name
- michaelaz774/MBrain★ 9same name
- Dragoon0x/product-skills★ 4same name
- afeezadebayo/thinking-decisions-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 6 edited versions we read keep these.
2
Distinctive
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 16
Start from the base rate of the reference class (the Historian supplies it), then adjust for specific, verifiable advantages only.
- Distinctive: line 22
Every memo names the single assumption that collapses the case if false (the Skeptic supplies it) and the cheapest, fastest experiment that …
“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Decision frameworks for the Chairman
Compact reference. Apply silently; cite a framework in the memo only when it changed the outcome.
1. Door type first
One-way door (hard to reverse: new entity, key hire, public launch, big contract, reputation) demands the full protocol and a pre-mortem. Two-way door (cheap to reverse) should be decided fast and small; over-deliberating two-way doors is itself a failure mode. State the door type in every brief.
2. Expected value with ranges, not points
Force three scenarios per option: P10 (bad), P50 (base), P90 (great), each with rough probability and payoff. A project attractive only in its P90 story is a lottery ticket ONLY if it lacks any of the three: a capped stake (item 4), a named advantage over the base rate (item 3), or dated milestones that update the odds (item 5). With all three present, a tail-heavy bet is the correct SHAPE of bet for an owner whose stated goal requires tail outcomes; without them it is grandiosity, priced at zero. Watch asymmetry: prefer bounded downside with unbounded upside; reject the inverse even when the median looks fine.
3. Outside view before inside view
Start from the base rate of the reference class (the Historian supplies it), then adjust for specific, verifiable advantages only. If the case requires being in the top 5% of the reference class, the memo must name the concrete asset that puts it there.
4. Bet sizing (Kelly-flavored, conservative)
Size positions to survive being wrong: never stake an amount whose loss forces shutting other healthy operations. For a portfolio of projects, prefer several capped, falsifiable pilots over one uncapped commitment, but beware the opposite trap: ten 5% bets is a plan to learn, not a plan to win. Winning requires eventual concentration on the survivor.
5. Load-bearing assumption and cheapest test
Every memo names the single assumption that collapses the case if false (the Skeptic supplies it) and the cheapest, fastest experiment that could falsify it. Pilots are falsification experiments with budgets, not "soft launches".
6. Tripwires beat willpower
Decisions degrade by drift, not by formal reversal. Predefine observable conditions (metric below X by date Y, regulatory event Z, competitor action) that automatically reopen the decision. A tripwire that triggered and was ignored is the most important finding any review can produce.
7. Opportunity cost is the silent option
"Do nothing" and "do this instead with the same resources" are always on the ballot. For an agency, internal engineer-months are priced at billable rates, and the founder's attention is the scarcest asset of all: a decision that consumes 20 founder-hours weekly competes with everything else he could compound.
8. Calibration loop
Predictions in memos are pre-registered with p = NN%, dated, and Brier-scored at review; unresolved predictions stay open and unscored. A position is judged VINDICATED or REFUTED only against its own pre-registered criteria — its PREDICTION and CHANGES MY MIND lines — never by post-hoc narrative fit. Believability weighting applies only at N ≥ 5 scored predictions; below that, no weights. The owner's overrides are scored by the same rules as the seats.
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.

