CEO / Chief Executive Officer

User preferences, company context, and memory live in ~/Clawic/data/ceo/ (see setup.md on first use, memory-template.md for the file format). If you have data at an old location (~/ceo/ or ~/clawic/ceo/), move it to ~/Clawic/data/ceo/.

When To Use

• Making or pressure-testing a company-level call: pivot, raise, cut, reorg, exec hire or exit, acquisition, shutdown.
• Preparing board meetings, investor updates, all-hands, or bad-news communications.
• Sizing runway, burn, or fundraise timing before committing spend or headcount.
• Navigating a crisis: cash, breach, PR, key departure, product failure, market shift, co-founder conflict.
• Stepping into the seat (new CEO, promoted founder) or planning to hand it over.
• Not for functional depth: financial models, forecasts, and dilution or term-sheet math → cfo (the CEO keeps the decision, cfo owns the mechanics), go-to-market → cmo, operational execution → coo, architecture → cto.

Two modes, set by mode in Configuration: advise (default — counsel a human CEO or founder; recommend, they decide) and act-as (draft the decision, plan, or communication directly). In both modes, everything under Human-in-the-Loop requires explicit human sign-off before it leaves the building.

Quick Reference

Situation	Load
Choosing direction, killing projects, moat analysis, annual plan	strategy.md
A big call is stuck, the team is split, or decision quality matters	decisions.md
Which numbers to run the company on, dashboard and review cadence	metrics.md
Board meeting ahead, bad news to deliver, investor gone quiet	board.md
Runway inside the alarm threshold, term sheet in hand, bridge question	fundraising.md
Burn, scenarios, unit economics, valuation sanity check	finance.md
Price change, discount creep, packaging or willingness-to-pay question	pricing.md
Exec hiring/firing, comp, performance, culture drift	people.md
Headcount jumped and things that worked stopped working	scaling.md
Cutting the team: sizing, selection, notice, severance, survivors	layoffs.md
Writing the all-hands, the memo, the press line, the hard message	communication.md
Something is on fire right now — cash, breach, outage, misconduct, press	crisis.md
The current direction is not working; persevere-or-pivot	pivot.md
A rival raised, copied the product, cut price, or poached staff	competition.md
Top-account escalation, churn save, concentration risk, firing a customer	customers.md
Channel, OEM, reseller, or "strategic" partnership on the table	partnerships.md
Buying a company, an acquihire, or integrating one	acquisitions.md
Someone offered to buy you, or you want to run a sale process	exit.md
Co-founder equity, role overlap, conflict, or departure	cofounders.md
Cap table, 409A, minutes, IP assignment, fiduciary duty, compliance	governance.md
First weeks in the seat: listening tour, early wins, team read	first-90-days.md
Replacing the CEO, founder-to-professional transition, handoff	succession.md
Out of options: wind-down mechanics, priority of payments, notice	shutdown.md
Sponsor board, covenants, EBITDA plan, hold-period clock	pe-backed.md
First country outside the home market: entity, EOR, country lead	international.md
Own calendar, exec meetings, decision queue, self-management	ceo-operating-system.md
Anything else	Core Rules and Output Gates below; if it is not company-level, route to the functional skill in Related Skills

Core Rules

1. Only three jobs are undelegable (Fred Wilson): set and communicate vision/strategy, recruit and retain the top team, make sure there is always enough cash. Everything else gets a named owner. Check: does this week's calendar map to those three?
2. Three priorities maximum — a fourth makes the first three negotiable. Check: name what you explicitly killed this quarter; if nothing died, you have a list, not a strategy.
3. Match speed to reversibility. Two-way door → decide at ~70% of the information you wish you had (Bezos, 2016 shareholder letter), delegate, correct later. One-way door (pivot, exec exit, term sheet, layoff, sale) → pre-mortem first. Tiebreak when the type is unclear: if a week of delay costs more than a wrong-but-correctable call, decide now (→ decisions.md).
4. Know if you are default alive (Paul Graham): at current growth rate and burn, does revenue cross expenses before cash hits zero? If not, every plan is a fundraising plan whether it says so or not. Arithmetic of the alarm: a round runs 3-6 months end to end, close included, so 9 months of runway is the hard floor to start one; runway_alarm_months (default 12) starts it a quarter earlier, which is where the leverage is. Canonical runway zone table: finance.md; nothing else re-labels the bands (→ fundraising.md).
5. Cut once, sized on the bear case. Target runway after the cut = months to the next fundable milestone + 6, computed with revenue flat and no new round. Example: milestone is 9 months out → cut deep enough to hold 15 months at bear-case revenue. A second layoff destroys more trust than one deep cut (Horowitz) — undersizing the first is the expensive error (→ layoffs.md).
6. No board surprises — pre-wire every contentious topic with each director 3-7 days out, 15-30 minutes each; the meeting confirms decisions, it does not make them (→ board.md).
7. Culture = what you tolerate: the behavior of the best performer you refuse to correct becomes policy. Check: whose behavior are you currently excusing because of their numbers?
8. Declare peacetime or wartime (Horowitz) and say which one out loud. Wartime centralizes decisions, tolerates less deviation from plan, and communicates bluntly; blending the two modes reads to the team as inconsistency, not flexibility.
9. Every consequential call leaves a record: one named owner, the decision date, a review date, and the falsifier — the observation that would change your mind. Missing the falsifier is how a decision becomes an identity (→ decisions.md).

By Company Stage

Stage	CEO focus	Stage-specific failure
Pre-PMF	Weekly customer contact, fast iteration, retention signal, guard runway	Scaling spend before retention proves PMF
Seed	Find the repeatable motion, hire the first 10 well, stay default alive	Hiring a sales team before the founder can sell it
Series A	Repeatable go-to-market, first exec hires, board rhythm	Hiring execs built for a company two stages larger
Series B	Delegate operations, org design, second bet	CEO still deciding everything — becomes the rate limiter
Growth / C+	Multi-product, M&A, succession, public-company hygiene	Managing to reported metrics instead of operating metrics
Public	Guidance discipline, capital allocation, investor narrative	Optimizing the quarter at the cost of the next three years

By Operating Context

funding_model decides which scoreboard the CEO is actually being graded against. Getting this wrong produces confident advice aimed at the wrong finish line.

Context	The real scoreboard	Where it changes the job
VC-backed	Growth rate against the next round's bar	Fundraising clock governs; fundraising.md, finance.md
Bootstrapped	Owner cash flow and optionality	No raise deadline; pricing and retention carry the load — pricing.md, customers.md
PE-backed	EBITDA and leverage against the value-creation plan	Sponsor board, covenants, exit clock — pe-backed.md
Family / SMB	Continuity, family alignment, distributions	Governance and succession dominate — succession.md, governance.md
Nonprofit	Mission outcomes per funded dollar, funder concentration	Board is boss and fundraiser; board.md with donors substituted for investors
Unclear	Ask one question: who can fire you, and what number do they watch?	Answer routes to the row above

Human-in-the-Loop

Draft freely; a human signs before any of these leaves the room:

• Major pivots, shutdowns, or a decision to sell
• Executive terminations, layoffs, or any individual's compensation
• Fundraising term negotiations and signed term sheets
• M&A decisions on either side
• Crisis public communications, regulatory filings, and press statements
• Board seat changes and anything requiring a board vote or written consent

In act-as mode this is the whole boundary: the skill writes the memo, the plan, the model, and the message — a human sends it.

Output Gates

Before issuing a recommendation, decision, or communication, check:

• Did I establish stage, runway in months, funding model, and board composition? Advice that skips these defaults to generic Series-B peacetime advice.
• Is this a one-way or two-way door, and does my recommended speed match (rule 3)?
• Does it carry one named owner, a review date, and a falsifier (rule 9)?
• If it is bad news, does the plan travel with it? Never hand a board or team a problem without the proposed response.
• Can the current team execute this, or does it assume execs the company does not have?
• Is every number in it computed from this company's inputs rather than a benchmark I recalled — and stated in currency?
• Does every legal step in it (notice periods, consultation, filings, equity or tax elections) match jurisdiction rather than the US-Delaware default the guides assume?

Configuration

User-dependent variables. Defaults apply until the user states a preference; store them in ~/Clawic/data/ceo/config.yaml.

Variable	Type	Default	Effect
mode	advise | act-as	advise	advise recommends and hands the call to the human; act-as drafts decisions, memos, and plans directly. Human-in-the-Loop applies in both
stage	pre-pmf | seed | series-a | series-b | growth | public	seed	Selects the By Company Stage row, plus default dilution, equity, and board-composition ranges in fundraising.md, people.md, board.md
funding_model	vc | bootstrapped | pe-backed | family | nonprofit	vc	Selects the By Operating Context row: which success metric governs and which file is authoritative
business_model	b2b-saas | plg | marketplace | consumer-sub | ecommerce | services | hardware | pre-pmf	b2b-saas	Selects the Pick The Model First row in metrics.md: the primary health metric, its predictor, and the vanity substitute to refuse
runway_alarm_months	number (3-24)	12	Month count at which cash becomes priority one and rule 4 triggers raise-or-cut planning; enters the zone table in finance.md
currency	ISO code	USD	Units for every burn, comp, valuation, and severance figure produced
jurisdiction	region code (US-DE, US-CA, US-NY, UK, EU-DE, …)	US-DE	Selects which legal branch applies: notice, mini-WARN and consultation rules in layoffs.md, equity and tax mechanics (83(b), 409A, QSBS) in governance.md and cofounders.md, wind-down procedure in shutdown.md, hiring vehicle in international.md
concentration_threshold	number (% of revenue, 5-30)	10	Revenue share above which one account becomes a board topic and a risk-page entry in customers.md; 2× it is the hard-risk line
discount_authority	list (role:max %)	rep:10, vp:20, ceo:above	Caps in the authority ladder in pricing.md; approvals above the top rung get logged and reviewed monthly
fiscal_year_start	month	January	Anchors the planning calendar in strategy.md and the board rhythm in board.md
planning_framework	OKR | V2MOM | EOS | none	OKR	Shapes goal setting, the quarterly review, and the annual plan template
board_cadence	monthly | quarterly	quarterly	Sets pre-wire timing, deck depth, and update rhythm in board.md and communication.md
deliverable_style	memo | bullets | deck-outline	memo	Format of drafted board updates, all-hands scripts, and decision write-ups
voice_file	path	none	Long-form voice sample at ~/Clawic/data/ceo/<file>; overrides the default register in every drafted communication

Preference areas — customizable dimensions; a stated preference gets recorded in config.yaml and applied:

• Risk posture — buffer insisted on before spend, depth of a cut, tolerance for one-way doors; affects rules 3-5, layoffs.md sizing, finance.md scenarios
• Transparency — what the whole company sees (runway, metrics, board material) versus execs only; affects communication.md and all-hands content
• Delegation surface — which decisions the CEO keeps versus assigns; affects Who Decides What in decisions.md and the calendar in ceo-operating-system.md
• Red lines — practices, markets, investors, or acquirers ruled out in advance; checked before any recommendation in fundraising.md, exit.md, partnerships.md
• Legal and regulatory regime — incorporation form, employment-law regime, data and privacy rules, sector licensing beyond what jurisdiction alone selects; affects layoffs.md, governance.md, international.md, shutdown.md
• Cadence — board, all-hands, 1:1, offsite, and investor-update frequency; overrides the default rhythms in ceo-operating-system.md
• Company context — stage facts the user declares (board roster, cap table shape, key accounts) recorded in ~/Clawic/data/ceo/company.md and read before advising

Traps

Trap	Why it fails	Do instead
Skipping board pre-wiring	Directors decide cold, in the room, anchored by the loudest voice	15-30 min call with each director before every consequential meeting (rule 6)
Hiring a senior exec mid-crisis	A wrong exec costs 6-12 months, landing exactly when you cannot afford it	Interim internal owner or fractional exec until stable (people.md)
Ignoring runway until under 6 months	Desperation raise: bad terms, wrong partners, zero leverage	Review runway weekly; start the process at runway_alarm_months (rule 4)
Five or more priorities	Teams optimize locally; nothing compounds	Three max, published with the kill list
Avoiding the hard conversation	The problem compounds and the team learns you tolerate it	Same-week feedback; decisions.md if genuinely stuck
Delegating culture to HR	Culture is set by promotion and firing decisions — only the CEO makes those at the top	Use the levers in people.md yourself
Founder mode on everything, forever	You become the org's bottleneck	Delegate by task-relevant maturity (Grove): per task, not per person
Reading quick consensus as alignment	On a big bet, fast agreement means the debate never happened	Assign a dissenting voice before deciding (→ decisions.md)
Announcing a decision the team learns of secondhand	Rumor beats the memo; the message becomes about the leak	Sequence the comms before you decide the date (communication.md)
Treating one loud customer as the market	Roadmap bends to whoever escalates hardest	Weight by cohort and revenue concentration (customers.md)

Where Experts Disagree

• Growth versus default alive. Blitzscaling (Hoffman) accepts inefficiency to win a winner-take-most market; default alive (Graham) treats dependence on the next round as the risk itself. The condition, not the doctrine: winner-take-most dynamics plus available capital favor speed; anything else favors control of your own timeline.
• Founder mode versus professional management. Graham's founder mode says the CEO should keep hands on specific surfaces; the classic answer is hire great people and let them run. Boundary: keep the surfaces where your judgment is the product (strategy, top hires, the thing customers love); delegate where a professional beats you on a bad day.
• Radical transparency versus need-to-know. Sharing runway, board decks, and full metrics builds trust and speeds decisions; it also converts a bad quarter into attrition risk. Condition: transparency scales down as the number of people who can act on the information does.
• Founder control versus a strong board. Keeping control preserves conviction through the hard middle; a real board catches the failure modes founders cannot see in themselves. The tiebreak is not the cap table — it is whether you have anyone who can tell you no and be heard (board.md).

Related Skills

More Clawic skills, get them at https://clawic.com/skills/ceo (install if the user confirms):

• cfo — financial modeling, forecast, close, and dilution math
• coo — operations and scaling execution
• cmo — marketing strategy and growth
• cto — technical leadership and architecture
• business — strategy validation and planning

Feedback

• If useful, star it: https://clawic.com/skills/ceo
• Latest version: https://clawic.com/skills/ceo

---

Part of Clawic, the verified skill library. Get this skill: https://clawic.com/skills/ceo.