Most rounds aren't lost on the pitch. They're lost on the process — the list you never built, the intro you didn't map, the follow-up that landed three days late. Fifteen skills for Claude Code and Codex run the raise as a system — starting with whether and when to raise at all — so your attention goes to the conversations that matter.
Inbound won't fill a round. You build the list, find the paths in, and keep every thread moving in parallel — without dropping one.
And once you're in conversations, the operational signals matter as much as the pitch: how fast you follow up, how specific your outreach is, how ready your data room is the moment someone asks. Run the skills when you need them; each uses your shared project files and adapts to a quiet, passive raise or a sharp, time-boxed process.
A raise is a parallel operation with a dozen live threads, each decaying if you miss the next action. Here's where it leaks.
You take the meetings that come instead of building the 40 funds whose thesis actually fits. The round never gets a real top of funnel.
You email partners with no warm path mapped. A trusted intro beats ten perfect cold emails — but only if you find it first.
The same email with the name swapped. A partner smells it in one line. Specificity is the difference between ignored and answered.
Twelve conversations, twelve next-actions, no system. The follow-up that comes three days late tells the investor how you'll run the company.
A partner asks for the room and you spend two days assembling it. Readiness on the spot is itself a signal — absence of it is too.
You go quiet on the funds that passed. They're often the leads of your next round — if they watch you execute for six months.
It starts with a decision, not a deck. You fill the context once; every skill inherits its numbers, names, and narrative. The phases run roughly in order, with warmth running the whole time and the close at the end.
Whether and when to raise — each founder weighs in, the verdict has teeth.
Set the source of truth, then choose how to run the raise.
Targets, paths in, the story, and the numbers behind it.
Reach out, manage the pipeline, pre-answer diligence.
Keep everyone warm, pace the whole thing, and drive a clean window to the wire.
An on-demand workflow for your raise: invoke a skill in Claude Code or Codex to prepare materials, update the project files, and surface the decisions that need you. Scheduling and external integrations are configured separately.
Prefer ordinary Claude chat? Copy Prompts copies the chat setup and that skill's prompt together. Paste into Claude, fill in the placeholders, and supply your context. Choose View prompts to read them first or copy just the skill prompt. See the full prompt sequence.
Each founder answers a short questionnaire independently — readiness, dilution tolerance, time appetite. It surfaces where you disagree, runs the dilution math and the alternatives, and returns a verdict with teeth: raise now, not yet (with the milestone that flips it), or don't raise.
Interviews you, pressure-tests every claim, and flags what's missing or weak before it ever reaches an investor. Everything downstream inherits this file.
Picks up the verdict from raise-decision and sets the mode — a quiet "always slightly raising" or a sharp, time-boxed process — writing the intensity into a plan file every downstream skill reads.
Ranked by fit, with evidence coverage and eligibility shown separately. See which partners match, which need more research, and which conflicts rule out outreach.
Maps the strongest warm intro through your network and drafts the forwardable note your connector can send in fifteen seconds.
Headlines that are claims, not labels. A traction page partners screenshot. Creates a deck with your app’s available tools, or returns clearly labelled slide content.
A driver-based model plus the one-page assumptions memo you actually defend. Runway and ask reconcile exactly with the deck.
Rehearse your opening, face skeptical investor questions one at a time, and tighten answers around evidence. Unknowns become preparation tasks.
A sourced brief on the fund, partner and relationship, with tailored talking points, likely objections, questions to ask and a proposed next step.
A researched hook, one proof point, a small specific ask — never the same body with the name swapped. Warm and cold variants per target.
A weekly Monday brief of what needs action, same-day follow-up generators, and stage tracking across every thread in parallel. Speed is the message.
The full index, a claim-to-proof map, a ranked gap list, and staged access. Ready the moment a partner asks — which is itself the signal.
Monthly update drafts when you run the skill, with an honest challenges line. Out-execute the reason they passed, in writing, and the next round writes itself.
Decides which skill runs next and why, keeps the plan, sheet, and pipeline files consistent with each other, and protects founder attention. The glue that orchestrates the whole raise.
A plain-language term-sheet readout — pool placement, preference, board — negotiation prep, and a real close window. Clarifies so you decide; not legal advice.
A suggested cadence for invoking the skills. Each uses your latest project files. For recurring runs, configure a scheduled task in your app; this package does not install one or connect your inbox.
| When | What runs |
|---|---|
| Before anything | The whether/when decision — every founder scores it independently |
| Per meeting | Investor brief and pitch practice before; specific follow-up after |
| Daily | Overdue-thread scan — anything live untouched >5 days |
| Weekly | Monday pipeline brief; the next wave of outreach |
| Monthly | Investor update to both tracks — committed and passed |
| Per stage | Decision gate — keep going, pause, or flip passive→active |
| At close | Term-sheet readout, negotiation prep, the close window |
Claude Code and Codex use the same fifteen skills and raise files. Pick one install method per app, open your raise folder, and start with the decision.
--app both option installs standalone skills for Claude Code and Codex in one raise folder. Use /raise-decision for standalone Claude skills and $raise-decision for Codex. Finish one app’s file edits before switching to the other. Shared-folder setup →The skills are free under MIT. Your AI app’s subscription or usage charges still apply. No email, CRM connection, or background schedule is installed.
Starting a raise? Follow the six-week plan, see the process and dependency maps, use the daily and weekly check-ins, or copy the Claude chat prompt sequence.
Treat six weeks as a target to test against evidence. This is a suggested operating plan for a prepared early-stage raise, not a promise or an industry conversion benchmark. Warm relationships, investor decision calendars, your instrument and legal readiness determine what fits. If you are starting cold or need an institutional lead, the date may need to move. Use existing work and enter at your actual stage.
Define the finish line first: the full target amount received, or an explicitly agreed first close with its own amount. A positive meeting, verbal interest, term sheet and signed investment documents are different milestones; none should be reported as cash received. Put a real start date, target funding date, amount, currency, raise lead and company jurisdiction in fundraising-plan.md. Ask counsel and prospective investors early whether the timetable is workable.
Begin with the copyable kickoff request, then use the daily and weekly pacing prompts throughout the six weeks.
raise-decision to record or confirm the team's decision, raise-context to reconcile facts, and fundraising-plan to set the active process, dates, ownership and fallback. Define the milestone the money funds and who protects company operations.deck and financial-projection for a defensible story, ask and cash plan. Use data-room to identify the essential evidence and assign missing items. Get counsel started on corporate, cap-table and instrument readiness now.investor-list and warm-intro-map to research a manageable target set. Around 30–40 qualified candidates is a planning starting point for an active process, not a quota. Separate likely leads from followers; exclude blocked targets and finish research before treating uncertain fit as qualified.pitch-rehearsal. Once the essential materials are credible, use outreach to draft a small first wave, for example 3–6 suitable investors. Review and send through your normal channel, then record actual sends in pipeline. Avoid spending a whole week polishing cosmetics.Friday checkpoint: consistent context, usable deck/model, core diligence evidence, a researched first wave and named owners. Some introductions or meetings should be in motion; if key facts or documents are missing, name the blocker and revise launch dates.
Refine the story from the first conversations, then approach the strongest qualified targets promptly. Use a short calibration wave to learn; do not hold your best investors back until the final weeks. Batch meeting slots to protect preparation and company work. Before each meeting run investor-brief; rehearse the hardest objection with pitch-rehearsal. Ask who decides, whether they can lead, their likely ticket, what evidence they need and their actual decision timetable.
After every meeting, supply factual notes to pipeline, draft the promised follow-up and send it when ready. Record the next step, owner and date, distinguishing a proposal from an investor-confirmed commitment. Follow agreed dates; do not send daily reminders merely because you check the pipeline daily.
Friday checkpoint: track substantive replies, completed meetings and dated next steps. Have second meetings or specific requests emerged? If meetings are not materializing, fix targeting, introduction paths or the message before expanding volume.
Prioritize investors taking concrete next steps: bringing another decision-maker, requesting relevant diligence, or agreeing a decision meeting. Use data-room to close evidence gaps and prepare appropriate access; you still control actual sharing. Reconcile every answer with raise-context and the model. Keep additional qualified conversations moving while the strongest prospects evaluate.
For a lead-dependent round, identify at least one plausible lead with an explicit process and dates. For a round that does not require a lead, track investor-by-investor decision and funding dates instead. Use closing to prepare questions and a checklist as terms emerge; use investor-comms for useful, factual updates to active investors when there is real news.
Friday checkpoint: a credible path to decisions should exist. No follow-on meetings, diligence or named decision timetable means the six-week target is at risk. Run pace and re-plan now, not in week 6.
Ask advanced investors what remains for a decision and arrange the relevant partner or decision-maker conversations. Use closing to compare proposed terms and prepare negotiations with counsel. Confirm amounts, conditions and each investor's process. Keep verbal interest, written indications and signed documents separate.
For a priced, lead-dependent round, use this week as a planning checkpoint for lead/terms progress and an agreed diligence/legal timetable. Counsel may require more time; this is not a universal two-week closing rule. For rolling investments, process each eligible close on the applicable documents and conditions. Respect any agreed exclusivity or other obligations.
Friday checkpoint: there should be a specific route from the current status to funding: outstanding decisions, diligence items, documents, approvals, owners and dates. Without that route, mark the deadline red and decide whether to extend it or pursue a feasible first close.
Run closing with the current documents and counsel's updates. Maintain an investor-by-investor checklist for remaining diligence, allocations, approvals, legal documents, signatures and expected funding dates. Resolve critical questions promptly; agree a delivery date when an answer needs work. Follow up with interested participants using real progress and genuine deadlines.
Keep a short daily blocker review with the responsible people. Confirm the administrative steps with counsel, investors and your finance owner. A spreadsheet of hoped-for checks is not a funded round.
Friday checkpoint: every amount needed for the planned close has a documented status, owner and expected funding date. Identify any gap explicitly; do not count unsigned or conditional amounts as received.
Work through the remaining approvals, documents, signatures and funding conditions with counsel. Track expected transfers and verify actual receipts against the company account with your finance owner. Use pipeline to keep signed and received amounts distinct. Close out the cap-table and document records with the responsible professionals, then use investor-comms to draft the appropriate investor update.
Finish-line checkpoint: report the amount actually received versus the original target, remaining commitments and conditions, and the resulting cash plan. If only a first close is complete, call it a first close and set the next milestone. Do not silently redefine a partially funded round as the full target achieved.
The week labels above are our planning recommendation. Background reading: YC on parallel investor conversations and YC's seed fundraising guide. Instrument and closing mechanics should be confirmed for your company with counsel.
Think of a shared workspace with specialised steps. Company facts live in the context, strategy lives in the plan, and investor activity lives in the pipeline. Each skill reads the relevant inputs and produces something the next step can use. You supply decisions and real-world outcomes; the package does not automatically execute the arrows.
This is the usual first-use order. Preparation branches can run alongside one another, and you revisit them when facts change. Diligence preparation begins before launch; meeting preparation repeats for every meeting. Week labels describe a six-week target, not a guaranteed timetable.
The repeating operating loop: investor-brief → founder meeting → factual notes into pipeline → follow-up draft → founder sends → pipeline records the actual outcome. Pitch rehearsal helps before the meeting; data-room supports diligence. Investor-comms keeps the appropriate people informed.
Each skill box names its main output. Solid arrows show a principal input/output handoff. Dashed arrows show helpful context or a review loop. Dotted arrows require a human action and a confirmed result. The shared-input line is shorthand, not a claim that every skill has identical prerequisites. Select a skill to jump to its exact input/output contract below.
The graph deliberately omits repeated connections to keep the main paths readable. For example, deck and model must reconcile; the data-room uses their evidence; an investor brief can inform rehearsal. The table gives the fuller dependencies and marks optional inputs. On a narrow screen, scroll the graph or use the text table.
These are workflow requirements, not a requirement to start every raise from an empty folder. Provide existing current outputs. In ordinary Claude chat, paste or attach the latest versions; return/save the results yourself when file tools are unavailable. A document being produced does not mean it was sent, published, signed or funded.
| Skill / first use | Reads or needs | Produces → used by |
|---|---|---|
| raise-decisionDecide · week 1 | Founder positions, objectives and rough financing facts. Existing raise-context is helpful, not required. | raise-decision.md: raise now / not yet / do not raise; disagreements, rationale, gates. Next: Context and plan; a not-yet/no verdict routes to milestones or an alternative. |
| raise-contextFoundation · week 1 + refreshes | Founder answers and supplied company evidence; an agreed decision if already recorded. Missing facts are interviewed. | raise-context.md: dated facts, narrative, raise parameters, networks and readiness gaps. Next: Shared company facts for every preparation and operating skill. |
| fundraising-planFoundation · week 1 + decision gates | Raise decision and current context, including runway and traction. Unresolved decision conflicts must be addressed. | fundraising-plan.md: mode, raise lead, window, intensity, ownership and decision gates. Next: Targeting, outreach, pipeline, data room, updates, closing and pace. |
| financial-projectionBuild · week 1 + actuals updates | Context: actual business drivers, cash, costs, unit economics and use of funds. Unknown assumptions must be surfaced. | financial-assumptions.md and a model; financial-model.xlsx when tools allow, otherwise labelled tables/formula map. Next: Deck ask/runway, data-room financial evidence and financing decisions. |
| deckBuild · week 1 + revisions | Current context and defensible claims. Model assumptions/results where needed for the ask and runway. | deck.md: slide narrative and evidence; editable presentation if available tools support it. Next: Pitch practice, meeting material and the data-room claim-to-proof review. |
| investor-listBuild · week 1 + selective expansion | Plan, context/ideal investor profile, conflict criteria and source-backed research. | investors.tsv and investor-scoring.md: partners, fit, evidence coverage, eligibility and gaps. Next: Introduction mapping, outreach and initial pipeline creation. |
| warm-intro-mapBuild / reach · before an approach | Investor list and each founder's actual network from context; verify inferred relationships. | Paths with relationship owners, connector requests and forwardable drafts; Warm Path? updates in investors.tsv. Next: Outreach through one supported path, or a personalised cold approach. |
| data-roomPrepare / diligence · week 1 onward | Plan and context; available deck claims, model and supporting documents for the evidence review. | Room index, claim-to-proof map, ranked gaps and access recommendations. Next: Founder-controlled sharing, diligence responses and closing preparation; no room hosting is bundled. |
| pitch-rehearsalPrepare / meetings · repeat as needed | Available context, deck, assumptions or supplied pitch. An investor brief is optional; missing files do not block practice. | Interactive feedback and revised answer outlines; pitch-rehearsal.md notes when requested or at session end. Next: Founder delivery and proposed deck improvements; it does not silently change company facts. |
| outreachReach · weeks 1–2 + qualified waves | Plan, context, investor list and warm-intro map; named sender and qualified target. | Partner-specific unsent messages, connector requests and follow-up proposals. Next: Founder reviews/sends through an authorized channel; actual outcomes are then supplied to pipeline. |
| pipelineRun · daily + after interactions | Existing pipeline first, or investor list to initialize; plan, context and actual interaction updates. | pipeline.md: thread stages, owners, last touch and next action; briefs and unsent follow-ups. Next: Meeting briefs, updates, closing and pace. A draft alone never advances a contact stage. |
| investor-briefMeetings · before each call | Available context, investor row, pipeline, deck and meeting details. Supplied notes suffice; browser/CRM/calendar sources are optional. | investor-brief-[fund-or-partner].md: meeting status, history, fit, talking points, objections and proposed next step. Next: Founder meeting preparation and optional rehearsal; actual meeting outcomes return to pipeline. |
| investor-commsSustain · useful milestones + monthly | Context, plan and pipeline/recipient status; current and prior metrics for comparisons. | Unsent updates for live/committed investors and a separate passed/not-now track. Next: Founder-approved communication; confirmed responses can update the pipeline. |
| closingClose · prepare early; terms through funding | Proposed terms, plan, context and pipeline; counsel and investor updates as decisions progress. | Term-sheet explanation, negotiation preparation, close window and checklist. Next: Founders decide; counsel/investors/finance handle execution. Confirmed signatures and cash update pipeline. |
| paceAcross all stages · weekly + when blocked | Whatever decision, context, plan, investor list and pipeline files exist. Missing or stale inputs are themselves findings. | Next-skill recommendation, consistency check and attention review; dated pacing log if requested. Next: Routes work and proposes re-planning. It does not run a scheduler, send messages or make investor decisions. |
raise-context.md, then reconcile the deck, model, data-room claims and affected drafts.pipeline; keep proposed next steps distinct from agreed dates.pace to diagnose it, then explicitly agree changes through fundraising-plan.Example: after an investor asks for retention evidence, pipeline records the request and its owner/date; data-room identifies the missing proof; the founder supplies a verified analysis; context/deck/model are reconciled where affected; an unsent reply is prepared; the founder sends it; pipeline records the real send. Pace checks whether the delay threatens the funding deadline.
Ready to run a step? Each skill name links to its copyable Claude prompt. Use the six-week plan for dates and the pacing routine to decide what runs next.
Yes. Paste these prompts into an ordinary Claude chat; no plugin or slash commands are required. They are concise, self-contained adaptations of the 15 workflows. The full installed skills contain additional templates and detailed instructions. You can optionally attach a skill's SKILL.md and its bundled references for that fuller version.
How to use the sequence: paste the setup once, then run one prompt at a time. Answer Claude's questions and review each output before proceeding. Replace bracketed placeholders. Keep the latest context, plan and investor table; attach or paste them when starting another chat. A filename below is a label for the output to save, not a claim that Claude can see or edit a file on your computer.
Week 1: steps 1–9 prepare the raise; run the model and deck together and revisit the deck after the numbers reconcile. Start step 10 as soon as the essentials are ready. Weeks 2–6: repeat 11–13 around meetings and reviews; use 14 for relevant updates and 15 as terms emerge. Start the closing checklist early. If you already have a current, agreed output, provide it and skip repeating that step. Do not run all prompts in one message.
Help me run my fundraise in this Claude chat. Use only the latest company facts and outputs I supply here or attach; do not assume access to other chats, my computer, inbox or CRM. Company/stage/jurisdiction: [details]. Amount/currency: [target]. Start/deadline: [dates]. Close means: [first close amount / full round received]. Raise lead: [name]. Ask for missing facts in small batches. Mark unknowns and distinguish evidence, assumptions and forecasts. If browsing is available and appropriate, cite sources and dates; otherwise say research is unverified. Treat messages as drafts. Do not send, share, book meetings or claim reminders are configured. Return each named output as copyable Markdown/TSV, or a downloadable file if available. Never claim it was saved locally unless that actually happened. At the end of each step, list the outputs I should keep, unresolved questions and the next prompt to run. I will provide the latest outputs when continuing in a new chat.
When: Week 1, or confirm an existing decision. Keep: raise-decision.md.
Help us decide whether to raise now, later, or not at all. Ask which founders are involved and collect each founder's view on readiness, dilution, time commitment, what capital unlocks and the alternative without funding. Surface disagreements without inventing consensus. Use supplied cash, runway and traction; ask for missing facts. Compare relevant funding alternatives and show dilution calculations with explicit assumptions. Give a reasoned recommendation for us to decide, measurable readiness gates if needed, and a revisit date. Return a dated raise-decision.md draft.
When: Week 1; refresh when facts change. Keep: raise-context.md.
Build our fundraising source of truth from the supplied files and answers. Interview me in batches of 3–5 questions about the company, founders, product, customer problem, traction and metric definitions, business model, market, competitors, differentiation, raise amount/instrument, use of funds, cash/runway, ideal investors, each founder's network and communication style. Separate actuals, forecasts and assumptions. Flag contradictory figures and mark unknowns TBD. Return a dated raise-context.md draft plus green/yellow/red readiness gaps and the most important issue to fix before outreach.
When: Week 1; revisit at decision gates. Keep: fundraising-plan.md.
Using our agreed raise decision and current context, plan a six-week raise. Ask for the start date, funding deadline, amount, stage, jurisdiction, current investor conversations, possible lead, raise owner and weekly capacity. Confirm whether close means first money received or the full round funded. Assess whether active, passive or hybrid fits; do not promise the deadline. Produce weekly milestones, owners, evidence needed, investor decision and legal dependencies, outreach waves, daily/weekly reviews and fallback decisions. Separate planned from investor-confirmed dates. Return fundraising-plan.md; surface any conflict with our decision or runway.
When: Week 1; revise from real feedback. Keep: deck.md.
Using our current context and plan, draft a concise investor pitch with a clear problem, customer, solution, why now, traction, business model, market, competition, team and funding ask tied to milestones. Give each slide a claim-based headline, supporting evidence, proposed visual and speaker notes. Cite the supplied source for material claims; flag missing proof and conflicting numbers. Keep forecasts labelled and remove invented urgency or certainty. Return deck.md. If I request slides, use available file tools and verify the result; otherwise provide slide-ready content without claiming a PowerPoint exists.
When: Week 1; update with actuals. Keep: financial-assumptions.md and model.
Build a driver-based fundraising model from our supplied actuals. Ask for missing cash, revenue drivers, costs, hiring plans and funding assumptions before calculating. Model 18–24 months of monthly revenue, costs, cash and runway, with base and downside cases. Separate actuals, forecasts and editable assumptions; reconcile the ask, use of funds and milestone with the deck. Show formula logic, key sensitivities and a one-page financial-assumptions.md memo. If spreadsheet tools are available, create a formula-based workbook and verify its calculations; otherwise give tables/CSV and a formula map, clearly labelled as a fallback.
When: Week 1; expand selectively. Keep: investors.tsv and investor-scoring.md.
Build or rescore investor targets for our agreed stage, sector, geography, desired individual check size and lead requirement. Size the list to our plan; do not pad it. Research funds AND partners using available browsing, citing sources and dates; without browsing, use supplied evidence and label it unverified. Gate documented mandate mismatches and direct competitor conflicts before ranking. Score known dimensions 0, 0.5 or 1 with weights: sector 25, stage 20, geography/check size/lead behavior/portfolio relevance/recent activity 10 each, partner relevance 5. Report fit = earned weight / known weight and coverage = known weight / applicable weight, both as percentages. Unknown is not zero or clearance; a sector-only match is 100 fit at 25% coverage and needs research. Exclude genuinely inapplicable dimensions from both denominators; if no evidence is known, report fit as unknown. Missing mandatory fit or conflict evidence means research needed, not eligible. Explain eligibility and research gaps. Return investors.tsv and investor-scoring.md, preserving existing IDs, notes and columns; suggest a qualified first wave.
When: Week 1; before each outreach wave. Keep: Warm-intro map and draft notes.
Using our investor list and each founder's supplied network, identify one best introduction path per qualified target and a backup if supported. Name the relationship owner and label each path verified, unverified or cold. Do not infer a real relationship from a shared employer or social connection alone. Hold blocked targets out of recommended outreach. Draft a short founder-to-connector request and a forwardable investor introduction grounded in what the connector actually knows. List relationships we must confirm and targets needing cold outreach. Return the map and proposed investor-sheet updates; preserve other columns and send nothing.
When: Week 1; deepen during weeks 3–5. Keep: Data-room index and gap list.
Review our supplied company materials and prepare an investor data-room index. Cover corporate/cap-table records, financing, financials, metrics, product, customers/contracts, team, intellectual property and other relevant diligence. Map material deck claims to actual supporting documents. For every missing or stale item, show its importance, owner and proposed due date. Recommend staged access appropriate to the investor's progress and flag sensitive material for review. Return an index, claim-to-proof map and ranked gap list. Do not claim a room exists, upload files, change access or share anything.
When: Before outreach and important meetings. Keep: Feedback and rehearsal notes.
Run an interactive investor pitch rehearsal using our supplied context and deck. Start by inviting my two-minute opening, then wait for me. Give focused feedback on clarity, evidence and the ask, correcting unsupported claims against the source context. Role-play a skeptical but constructive investor and ask one question at a time. Wait for each answer before critiquing it and suggesting a tighter, factual outline. Do not invent my answers or assess voice/body language from text. At the end, summarize what we actually practiced, evidence gaps and next practice priorities. If I request a preparation pack instead, label it preparation, not a completed rehearsal.
When: Weeks 1–2; later qualified waves. Keep: Unsent partner-specific drafts.
Draft personalized outreach for the qualified investors I identify, using our context, investor evidence and introduction map. Confirm the sender/relationship owner and whether each approach is warm or cold. Use one specific, sourced reason for this partner, one verified proof point and a small concrete ask. Flag research gaps or portfolio conflicts rather than inventing personalization. Include the subject, email, any connector note needed and a proposed follow-up date consistent with our plan. Avoid fabricated introductions, endorsements, attachments or competing offers. Label all messages unsent; do not send or change last-contact dates.
When: Before every investor meeting. Keep: investor-brief-[fund].md.
Prepare a concise brief for [fund / partner] using the supplied company context, investor research, relationship notes and meeting details. Separate confirmed details from proposed times. Summarize fund and partner fit with sources/dates, relevant investments, conflicts, unknowns, relationship owner, last actual contact and open promises. Do not attribute a firm investment personally to a partner without evidence. Give three tailored talking points, likely objections with honest answer outlines, questions to ask and a proposed next step. Flag deck/context discrepancies. Return investor-brief-[fund].md; do not claim calendar/CRM access, book anything or change pipeline stage.
When: Daily and after every interaction. Keep: pipeline.md and follow-up drafts.
Update our investor pipeline from the latest supplied table and factual interaction notes. If either is missing, ask for it. Preserve investor IDs, relationship owners, stages, actual last-contact dates and agreed follow-up dates unless new evidence changes them. Track sourced, intro requested, contacted, meeting, diligence, partner meeting, term sheet, closed and passed as appropriate. Show due/overdue actions with owner and date; exclude passed/closed threads from live chasing. Separate verbal interest, signed amounts and received money; unknown amounts stay unknown. Return pipeline.md, today's three priorities and requested unsent follow-up drafts. Drafting is not contact; do not invent inbox sync or sent messages.
When: Monday/Friday; whenever a milestone slips. Keep: Pacing review and pacing-log.md entry.
Review our dated fundraising plan, context and latest pipeline as of [today]. Compare this week's milestones with actual evidence and assess the funding deadline as on track, at risk, off track or unknown. Show days remaining, stale/missing updates, the main blocker, the next skills to run and three priorities with owners/dates. Check consistency across the deck, model and pipeline, and protect time for company operations. Look two weeks ahead at decision, diligence, legal and funding dependencies. Recommend continue, adjust or re-plan; never silently move the deadline. Return a dated entry for pacing-log.md. Do not claim reminders are scheduled.
When: When there is useful news; monthly thereafter. Keep: Unsent investor updates.
Using the supplied current metrics, prior update and pipeline, draft separate concise updates for live/committed investors and investors who passed or asked to stay informed. Include a one-line summary, consistent metrics versus the prior period, concrete wins, an honest challenge, specific asks and the next milestone. Explain any changed metric definition and label missing comparisons. Respect communication preferences and never imply a passed investor has re-engaged without evidence. Keep verbal, signed and funded statuses distinct. Give suggested recipients and the relationship owner, but label everything unsent and do not publish or change contact dates.
When: Prepare early; use as terms emerge, especially weeks 4–6. Keep: Terms comparison and closing checklist.
Using the supplied proposed terms, cap table, plan and current investor statuses, prepare a closing brief for our decisions with counsel. Explain valuation/cap, dilution assumptions, option pool, preferences, control, investor rights and other material clauses present. Flag jurisdiction-dependent or unclear points; do not invent missing terms or tell us what to sign. Compare supported alternatives and identify negotiation priorities. Build an investor-by-investor checklist for decisions, diligence, approvals, documents, signatures and expected funding, with owners and dates. Keep proposed, agreed, signed and received amounts distinct; verify the close against the stated first-close/full-round target. Do not announce a close or initiate transfers.
Keep using the six-week milestones and progress scorecard. Browsing, file creation and scheduled tasks depend on the tools available in your chat; copying a prompt does not enable those capabilities.
Install the skills and open a private raise folder. Supply your latest deck, metrics, cash/runway, cap table and existing investor notes. Start with the request below; replace the brackets with real facts. Do not upload private company information to this public website.
Use the Fundraising OS skills to plan my six-week raise. Company: [name]. Stage and jurisdiction: [details]. Target: [amount and currency]. Start: [date]. Funding deadline: [date]. Close means: [full round received / first close of amount]. Current cash and runway: [facts or unknown]. Existing investor conversations and potential lead: [details]. Raise lead and available weekly time: [name, hours]. Review the files in this project. Confirm the raise decision, reconcile context, then use fundraising-plan to assess whether this deadline is credible. Save the dated weekly milestones, owners, readiness gaps and fallback decisions in fundraising-plan.md. Use pipeline for investor threads. Set up a pacing-log.md for our check-ins and keep the plan as the source for milestone dates. Give me the first three actions. Ask for missing facts; never invent them. Draft messages only. Do not claim that reminders or integrations are enabled.
For a specific skill, Claude plugin commands use /fundraising-os:pace, standalone Claude uses /pace, and Codex uses $pace. Substitute the relevant skill name. Natural-language requests can also name the skills. The pacing log is a project note you ask the assistant to maintain, not a separate installed service.
Use one owner, one plan and one pipeline. The raise lead owns the process. Every investor thread has a named relationship owner. Keep using the same project files in either app and finish one app's edits before switching.
pipeline for today's due actions, investor promises and overdue threads. Choose three actions that advance decisions. Respect existing follow-up dates.investor-brief, then rehearse one difficult question. Know the next step you want to agree.pace to compare this week's milestone with the evidence, choose the next skills, assign owners and protect time for running the company.These are suggested review slots, in addition to research, meetings and follow-through. Choose times that fit your workload; scheduling a review does not make the underlying work disappear.
Use pipeline for today's triage, then pace to assess our six-week plan. Today: [date]. Since the last check-in: [actual meetings, sent messages, replies, decisions, documents and money received]. Read fundraising-plan.md and pipeline.md. Preserve prior facts and agreed follow-up dates. Mark missing updates as unknown; a draft is not a sent message. Give me the three highest-priority actions, each with owner and due date; the biggest blocker; days left; and whether the funding deadline is on track, at risk, off track or unknown, with evidence. Append a dated check-in to pacing-log.md. Propose milestone changes explicitly rather than silently moving the deadline. Draft only; do not send messages.
Use pace for our weekly fundraising review, with pipeline supplying thread status. Compare this week's dated milestone with actual evidence. Show changes since the last review, stage movements, overdue promises, missing evidence and the next two weeks' critical dependencies. Keep verbal interest, written indications, signed amounts and received cash separate. Show the funding gap without guessing unknown amounts. Recommend continue, adjust or re-plan; explain why. Give next week's three priorities, owners and dates. Append the review to pacing-log.md and record any explicitly agreed plan changes in fundraising-plan.md.
Use a small scorecard with this week's plan, actual, change since last review, owner and next date. Count unique investor firms consistently and separate lead-capable firms from followers. Track:
Green: milestone evidence exists and the remaining decisions, documents and funding steps fit the confirmed timetable. Amber: an important step slipped or timing depends on an unconfirmed decision. Red: the required decision path is absent or a critical step cannot fit the remaining time. Unknown: updates are too stale or incomplete to assess. These are working assessments, not probabilities of success.
Agree activity targets from your actual starting pipeline and founder capacity. Review conversion using observed counts and comparable periods; do not treat a tiny sample as a forecast or assume a fixed number of emails guarantees funding. A week full of meetings without agreed next steps can still be off track.
Run pace immediately with the latest facts. Diagnose where progress stops: no replies suggests a targeting, access or message problem; first meetings with no next step suggest an unresolved fit or evidence issue; diligence without decisions calls for a clearer process, owner and date. Ask investors for specific feedback instead of guessing their motives.
If your round requires a lead and none has a credible decision path by week 3, mark the deadline at risk. If week 4 ends without actionable terms progress and a workable legal/funding timetable, treat a full close by week 6 as off track unless concrete evidence supports an exception. Those are planning escalation points, not universal fundraising laws.
Use fundraising-plan to compare an extended process, a feasible first close, a revised scope or a pause to improve evidence. Discuss financing changes with counsel and revisit the cash plan using actual runway. A smaller first close only helps if it funds a useful next milestone; it must not quietly replace the original target. If the round does not need a lead, assess signed participation and funding dates instead.
Record the decision, owner and revisit date. Do not manufacture competing offers, claim signatures you do not have, or impose a fictional investor deadline to rescue the calendar.
The skills run when you invoke them. For the manual version, put a short weekday check-in and Monday/Friday reviews in your calendar, open the same raise folder, supply the latest changes and use the prompts above. The saved plan and pacing log let the next session compare commitments with what happened.
If your host app supports scheduled tasks with access to this project, configure them separately using the daily or weekly prompt. Specify the timezone, review times, project path, start/end dates and where the result should appear. Verify that the task exists and can read the files; run a test and inspect its output. A normal calendar reminder can simply ask you to open the project when automated access is unavailable.
A scheduled review still needs current facts. Without connected, authorized inbox/CRM sources, you must supply updates. Missing updates should produce a stale-data warning, not invented progress. Creating a prompt does not create a reminder, send an investor email or sync an account. Stop or revise the schedule when the raise ends or the plan changes.
Yes. The same fifteen skills work in Claude Code and Codex. Claude plugin commands use /fundraising-os:raise-decision; standalone Claude skills use /raise-decision; Codex uses $raise-decision. Both local apps can read the same raise folder. Claude.ai web uses the instructions through guided chat.
The skills run when invoked. Daily, weekly and monthly cadences are recommendations; recurring runs need a separately configured scheduled task in your app. No email, calendar, CRM or Drive integration is bundled. The skills draft messages; sending or sharing requires your instruction and a connected tool. A draft does not count as a sent email.
When your app can generate files, the skills use its presentation, spreadsheet or code tools to produce them. They do not require a separately named “pptx” or “xlsx” skill. In a session without those capabilities, you get clearly labelled slide content or a text/CSV model with assumptions, rather than a claim that a finished file was created.
Passive is "always slightly raising": a rolling instrument kept quietly open, warm inbound only, founder stays building — lower intensity, no public in-market signal, but weaker competitive tension. Active is a structured, time-boxed process: full list, parallel conversations, an engineered close window — better terms and speed, but it consumes the founder and a stalled process is visible. Hybrid — passive posture with active bursts — is the most common reality. The fundraising-plan skill makes that call deliberately, and then configures every other skill to match the mode you chose.
Yes — that's the point of raise-decision. Its three outcomes are raise now, not yet (with the specific milestones that flip the answer and a revisit date), and don't raise (with what to do instead — revenue, angels, debt, or staying bootstrapped). No path silently defaults to "raise".
Yes. raise-decision collects each founder's position independently and puts your disagreements on the table before you commit. Downstream, the pipeline tracks an owner per thread and outreach drafts in the voice of whoever owns the relationship. Share one project folder (a private repo or synced drive) so the team reads the same context, plan, and pipeline files. Coordinate edits; the package does not merge simultaneous changes.
You can ask Codex to install the project skills, use the Claude Code plugin, or try the instructions in a Claude.ai project. The setup guide shows the exact commands. You supply the company facts and make the decisions.
Founders and small founding teams — deciding whether to raise, or running an early raise (pre-seed through Series A) — who can't rely on inbound and want the operational process handled with discipline — so their attention goes to the conversations, not the spreadsheet.
It's the companion OS. Atomic Scaling OS runs your growth operating system on the 3P3R Method®; the Fundraising OS runs the raise. Both turn repeatable work into skills while you make the decisions, under the same MIT license.
No. It explains term-sheet clauses in plain language and flags what's standard vs. off-market so you can decide and brief your lawyer efficiently. It never tells you what to sign — anything non-standard goes to a real startup lawyer.
Yes — MIT license, fork freely. Adapt the skills to your stage, sector, and geography, change the scoring, add your own. Pull requests welcome.
Serial entrepreneur and author of Atomic Scaling, building a portfolio of ventures across AI, wealth, and creative work. The Fundraising OS encodes the operational discipline of a raise — the list, the paths, the pipeline, the warmth — into skills any founder can run.