decision-memo
Skill4 starsChanged 2 months agoMITView on GitHub
Write a decision memo a decider can act on in one read: the situation, the options priced, the door type, one recommendation, and the dissent recorded. Use when a decision is stuck in threads and meetings, when a choice needs a sponsor's sign-off, when a team keeps re-deciding the same question, or when a reversible call is getting one-way-door treatment (or the reverse). Takes the question, the options, and the evidence on hand; returns the memo in the decision memo template and, once decided, the entry in the decision log.
5 other projects use a version of this.
Adopted by
Most recently taken up by
- Nomanriaz786/claude-consulting-teamtheir own version, first seen Nov 1, 2026
- Latifox/find-me-saastheir own version, first seen Sep 6, 2026
- Natan-Mohart/24-strategy-skills-for-claudetheir own version, first seen Aug 8, 2026
- XRAY-INC/claude-skills-for-leaderstheir own version, first seen Jul 13, 2026
- MaxKmet/idea-validation-agentstheir own version, first seen Apr 15, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- MaxKmet/idea-validation-agents★ 449same name
- Natan-Mohart/24-strategy-skills-for-claude★ 41same name
- XRAY-INC/claude-skills-for-leaders★ 27same name
- Latifox/find-me-saas★ 10same name
- Nomanriaz786/claude-consulting-team★ 1same 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.
In this repository37 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- ai-prdSkill
- competitive-intelSkill
- conductorSkill
- design-reviewSkill
- escalationSkill
- experiment-designerSkill
- feedback-synthesisSkill
- gtm-launch-plannerSkill
- launch-readinessSkill
- market-sizingSkill
- metrics-treeSkill
- okr-criticSkill
- persona-builderSkill
- pm-hiringSkill
- postmortem-facilitatorSkill
- pricing-packagingSkill
- product-analystSkill
- product-reviewSkill
- program-premortemSkill
- reg-gap-checkSkill
- roadmap-builderSkill
- evidence-researchSkill
- experiment-analysisSkill
- independent-reviewSkill
- lifecycle-conductorSkill
- portfolio-planningSkill
- product-definitionSkill
- release-governanceSkill
- spec-reviewSkill
- stakeholder-updateSkill
- story-writerSkill
- strategy-criticSkill
- user-interviewSkill
- write-prdSkill
- write-vision-strategySkill
The file
Files
Decision Memo: one page, one decider, one recorded answer
Decisions die in two ways. The reversible ones get committee treatment and take a quarter; the irreversible ones get made in a chat thread on a Friday. Either way the rationale evaporates, and the question comes back in six months to a room that has to guess why. This skill sizes the process to the door, forces the options to be priced, and puts the answer where the next person can find it.
Files this skill drives
- ../../templates/planning/decision-memo.md, the memo itself
- ../../templates/execution/decision-log.md, where the decided outcome lands as a numbered entry
- ../../frameworks/prioritization/decision-doors.md, the reversibility sheet that sets how much process the decision gets
- ../../frameworks/prioritization/weighted-decision-matrix.md, used only when three or more options and three or more criteria need comparing
- Reads: ../../templates/execution/stakeholder-map.md for who holds the decision. Structural technology decisions go to ../../templates/architecture/adr.md instead.
- Method background: one-way and two-way doors (Jeff Bezos, Amazon shareholder letter, 2015); the situation, complication, resolution spine (Barbara Minto, The Pyramid Principle, 1978), indexed in ../../knowledge/INDEX.md. Both explained here in this repository's own words.
When to use
- A question has been debated for more than ten minutes twice, by the same people
- A sponsor or forum must decide, and needs something shorter than the thread
- A team is about to make a one-way-door call informally, or is treating a two-way door as one
- An earlier decision is being reopened and nobody can find the original rationale
Inputs
The question in one sentence, the options already on the table, the evidence and the constraints (dates, budget, contracts), and the name of the person who holds the decision. Ask for these when missing: the needed-by date and what expires after it; the decider's name (if no one holds the decision, that gap is the first finding, and the escalation skill owns it); the "do nothing" option, priced honestly; and who disagrees and why. A memo with no dissent recorded usually means nobody was asked.
Workflow
1. State the decision as a decision
One sentence, in the form "whether to X or Y by date Z". "Discuss pricing" is a topic; "whether to move the pilot cohort to usage pricing before the renewal window closes" is a decision. If the sentence needs two "whether"s, it is two memos.
2. Classify the door
Score reversibility on the decision doors sheet: the cost to reverse (money, time, trust), the time until reversal becomes impossible, and who is affected by a reversal. Decision rule: a two-way door (cheap to reverse, reversal possible for months) gets a one-page memo, one decider, and an answer within two business days of the memo landing. A one-way door (expensive or impossible to reverse) gets the full memo, a review by the affected owners, and an explicit "what would make us wrong" section. Most decisions are two-way doors misfiled as one-way; say which this is and why.
3. Write the situation and the complication
Situation: two or three sentences of facts everyone already agrees on. Complication: what changed, or what the deadline forces. No adjectives, no options yet. If the reader would dispute a sentence here, it belongs in the evidence list with a link, not in the situation.
4. Price the options
Two to four options, always including "do nothing" and "decide later", each priced. For each: what it costs, what it gives up, the biggest risk, and the evidence that supports it. Decision rule: use the weighted decision matrix only when there are three or more options and three or more criteria that pull in different directions; run its sensitivity check (change one weight, does the winner change) and report the result. For two options, prose beats a matrix. A single option is a demand wearing a memo's clothing.
5. Recommend one, and record the dissent
One option, committed, with the reason it beats the runner-up and what is being given up. Then the dissent section: who disagrees, their strongest argument stated in words they would accept, and what evidence would change the recommendation. Dissent is recorded, never resolved by the memo; the decider weighs it. For a one-way door, add the reversal plan: the signal that would show the decision was wrong, who watches it, and what the fallback costs.
6. Get the answer and log it
The decider writes granted, declined, or deferred on the memo, with a date. Within a day, create the decision log entry: decider, options that lost, rationale including what was given up, who was told. A deferred decision gets a new needed-by date, or becomes a risk register row with the decider recorded as accepting the risk.
Output format
The memo, in this order and under these headings:
- Decision: one sentence, needed by [date], decider [name]
- Door type: one-way or two-way, with the reversal cost in one line
- Situation, then complication
- Options table: | Option | Cost | What it gives up | Biggest risk | Evidence |
- Recommendation and rationale, one paragraph
- Dissent: who, their argument, what would change the recommendation
- For one-way doors: reversal signal, watcher, fallback cost
- Outcome line for the decider: GRANTED / DECLINED / DEFERRED, name, date
Then the decision log entry, in that template's block.
Failure modes this skill guards against
- Committee treatment for a two-way door. A reversible call that took a quarter cost more than any wrong answer would have.
- Chat-thread treatment for a one-way door. Irreversible, undocumented, decided by whoever was online.
- The single option. A recommendation with nothing to compare against is a request for a rubber stamp.
- "Do nothing" left unpriced. It is always an option, and it is never free.
- Dissent smoothed away. Memos that read as unanimous produce decisions that get relitigated.
- The recommendation that hedges. "Option A or B, depending" transfers the analysis to the decider.
- Rationale reconstructed later. Logged within a day, or it is a rationalization.
Exit gate
The memo feeds ../../templates/execution/decision-log.md, which every gate in ../../os/STAGE-GATES.md reads. The memo is not done when it is sent; it is done when the outcome line carries a name and a date and the log entry exists.
What checks this
Some checks on this skill run automatically and some depend on a person. This section says which is which. It is generated by python3 tools/what_checks.py from the declarations it links to, and CI fails when it falls out of date.
- Checked by CI on every commit. This file must carry the seven sections every prose skill carries:
python3 tools/skill_rubric.py --min 7fails theskill-rubricgate inpython3 tools/ci_gate.pywhen one is missing. That checks structure, not whether the advice is right. - Checked by the runtime. This skill declares no gate in SKILL.graph.yml, because the document it produces is revisited on a cadence rather than approved once: it is an input to whichever gate is open, not an artifact any one gate holds. Nothing it produces joins a gate's approval manifest, so editing that document stales no gate. Each gate is recorded with
pmos gate --bank-id <bank>, which refuses the proof unless every question in that bank has been accepted and the bank is the current one, requires the proof to name itssource,source_sha256,actor_id,requester_id,decisionandapproved_at, and rejects self-approval and any approver the bank does not pin. Approver ids are typed, not authenticated, so a recorded approval is a local attestation, not proof of who signed. If the proof's source changes after approval the gate goes stale and nothing later completes until it is proved again. The journey run shows all six gates and a stale gate being proved again. - Checked by a person. The runtime checks who approved and that the evidence has not changed. It does not check whether the reasoning in this skill's Workflow was followed well; that judgment belongs to the people who sign the gate. The product owner signs all six gates; each gate's other sign-offs are listed in os/STAGE-GATES.md.
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.

