# /cold-start-kickoff — Internal kickoff playbook generator
You are the operations lead turning a Phase-1 niche commitment into a working manual. The audience is the Conversint cofounders (Patrick Saul + Daniel Nfodjo). The output answers every "what do I do Monday morning?" question without ambiguity. **If this doc has to be re-read with questions remaining, it failed.**
**Shared conventions:** Before planning, read `references/skill-operating-standards.md` and `references/output-paths.md`. This skill is an internal research/GTM artifact skill: it consumes upstream deliverables, writes to the canonical internal research tree, preserves source citations for external claims, and routes CRM/project follow-ups through the owning skills instead of writing those systems directly.
Load org context first
**Before any planning work, load every structural org-bundle file** (see AGENTS.md § Org Context Bundle for the authoritative current set):
- `references/org/identity.md`
- `references/org/pricing.md`
- `references/org/founders.md`
- `references/org/constraints.md`
- `references/org/positioning.md`
- `references/org/anti-patterns.md`
- `references/org/value-signals.md`
After loading, print one line: `Loaded org context as of <date>` (date = most recent commit timestamp on any file in `references/org/`).
**Critical pre-flight check:** per `references/org/positioning.md` § Current focus state, **Conversint may still be in the niche-discovery phase.** If no niche is committed, this skill should NOT generate an operational playbook — it should produce an exploratory planning artifact and label it as such. Ask the user: *"Is there a committed niche to operationalize, or is this still exploratory niche evaluation?"* before proceeding. If exploratory, scope the playbook deliverable accordingly (label as draft, mark all numbers as illustrative, do not commit weekly cadences).
**Capacity rail enforcement:** combined ~20–30 hrs/wk total founder time per `references/org/constraints.md`. Refuse to generate a playbook that exceeds this rail over the chosen horizon. The earlier 36+36 hrs/wk assumption is obsolete.
**Source citations:** Every external-fact claim in the playbook (tool-pricing references, partner-program eligibility criteria, CAC benchmarks, regulatory facts pulled from upstream research) follows the four-step rule in `references/client-deliverable-standards.md` §4 — capture URLs as you research or pull from upstream artifacts, cite inline at each restatement, list every source in a Sources section, and re-verify URL ↔ claim pairings before committing.
This skill is the operational counterpart to `/campaign-planner`. Where `/campaign-planner` produces one niche's GTM plan, this skill takes a *committed* niche pair (or trio) and produces a *cross-niche operations manual* with capacity math, labor split, and week-by-week execution.
The reference precedent is `research-thought-leadership/_inventory/cold-start-kickoff-playbook.md` (committed 2026-05-04). Read it before generating a new one — it's the canonical structure.
When to use vs. not to use
**Use this skill when:** - A Phase-1 niche commitment has been made (typically 1–3 niches) - Each committed niche has completed `/market-research`, `/research-usecase`, and `/campaign-planner` deliverables - The user explicitly wants an operations doc / kickoff manual / "what do we do Monday" plan
**Do not use this skill when:** - Niches are still being screened (use `/market-research` instead) - The user wants a single-niche GTM plan only (use `/campaign-planner`) - The user wants client-facing material (this skill is internal-only — see `/scope-engagement` or `/research-usecase` for client work)
Inputs
Acceptable input forms:
- **Comma-separated niche slugs** that exist under `research-thought-leadership/<slug>/` — preferred. Example: `smb-saas-soc2-hipaa-lite,ai-grant-writing-ops`.
- **Path to the triangle-comparison.md** under `research-thought-leadership/_inventory/` — the skill will infer Phase-1 niches from the "Active deep-dive list" table and the synthesis recommendation.
- **Free-text niche commitment** describing 1–3 niches to operationalize, where the user has named the niches but not the slugs. The skill will resolve slugs from existing folders or stop and ask.
Required upstream artifacts (per niche)
For each niche in scope, all three files must exist and be readable. If any are missing, stop and tell the user which one to generate first via the corresponding skill.
| File | Required for | Generated by | |---|---|---| | `research-thought-leadership/<slug>/market-research.md` | Verdict, CAC, time-to-first-deal, pricing, competitive set | `/market-research` | | `research-thought-leadership/<slug>/research-usecase.md` | Tool stack picks, partner-program selection, productization plan | `/research-usecase` | | `research-thought-leadership/<slug>/campaign-plan.md` | Channel plan, kill criteria, week-by-week motion, budget per niche | `/campaign-planner` |
Also recommended (not required):
- `research-thought-leadership/_inventory/triangle-comparison.md` — for cross-niche capacity math context
Scoping questions (one batched turn)
Before writing, confirm in a single `AskUserQuestion` turn (skip any question the input or available files already answer):
- **Niches in scope** — confirm the 1–3 niches and their slug folders. If the user gave free-text, present the resolved slug list for confirmation.
- **Phase-1 horizon** — 90 days (default) or other (60 days for fast-test scenarios, 180 days for slower-cycle niches like nonprofit/ministry).
- **Founder roster + capacity** — names, weekly hours per founder. Default: Patrick Saul + Daniel Nfodjo at 36 hrs/wk each (see precedent doc §4). Adjust if the user specifies different.
- **Cash budget envelope** — default $4,500–5,000 over 90 days for the precedent two-niche concurrent run; ask if the user wants a different cap.
- **Labor split preference** — default to "one founder leads each niche, both contribute outbound + content"; ask if the user wants a different split.
- **Output filename** — default to `cold-start-kickoff-playbook.md` for the first one. If a kickoff doc with that name already exists, default to `cold-start-kickoff-playbook-<YYYY-MM-DD>.md` to avoid silent overwrite.
If essential inputs are missing (e.g., a niche has no `campaign-plan.md`), stop here and tell the user which upstream skill to run first.
Eight-section workflow
Each section produces part of the final markdown. Walk them in order. Do not skip ahead.
Section 0 — Purpose + Phase-1 commitment summary
State explicitly:
- Owners (founder names)
- Phase-1 niches with one-line wedge each (read from each `market-research.md` exec summary)
- Phase-2 deferred niches (if any) and the activation trigger
- Three decisions to lock at the kickoff meeting (the most important ones from §15 below)
Section 1 — Target outcomes by week 12 (or chosen horizon)
Build a metrics table with floor / target / stretch tiers. Pull numbers from each campaign-plan's week-12 targets. Combine across niches.
Required rows:
- SOWs / retainers signed per niche
- Proposals out per niche
- Reply rate per niche (rolling 4-week)
- Discovery calls completed per niche
- Outbound touches sent (cumulative)
- Anchor essays published
- LinkedIn organic posts
- Cash burned
- Cash collected (deposits)
- Niche-specific milestones (partner badges, case studies, speaking slots — pull from each campaign-plan)
End the section with a 6-month milestone target (typically MRR run rate) — the "did the cold-start work" goal.
Section 2 — The hard economics
For each niche, a table pulling from its `market-research.md`:
- First-deal CAC range
- Time to first deal range
- Sprint price OR retainer price
- Deposit terms
- Gross margin
- Post-partner-badge CAC (if applicable)
- Upsell / extension pricing (e.g., Type II retainer, vCISO-lite)
Then **combined founder-time math**:
- Total 2-founder capacity over 90 days = ~1,440 hours
- Niche A plan hours (from campaign-plan)
- Niche B plan hours (from campaign-plan)
- Total = X hrs = Y% utilization
- Slack remaining for delivery on first wins
If utilization >90%, flag it explicitly as a risk and recommend cutting one niche or extending horizon.
Section 3 — The cold-start principles (always 4)
These are non-negotiable. Adapt the wording per niche but keep the four principles:
- **Trigger events beat broadcast.** Define the trigger event for each niche from its campaign-plan.
- **N hyper-targeted touches beats 10× generic.** Pick a defensible weekly volume (typically 80 across two founders; adjust per capacity).
- **Borrow credibility before you have your own.** List the credibility surfaces (partner badges, audit-firm referrals, speaking slots, anchor essays) for each niche.
- **First 1–2 deals are loss leaders.** Define the case-study or testimonial-discount terms per niche.
Section 4 — Labor split
Recommend a split based on niche characteristics. Justify each assignment:
- Niche-2 (technical-credibility-heavy, e.g., compliance, security) → assign to whichever founder has more eng/security background
- Niche-3 (relationship + writing heavy, e.g., grant ops, consulting) → assign to whichever founder has more nonprofit/relationship background
For each founder, produce a weekly hours table showing how their 36 hrs/wk break down:
- Lead-niche outbound (typical 8 hrs)
- Discovery + proposal calls (typical 6 hrs)
- Content + speaking + partner mgmt (varies)
- Other-niche outbound assist (typical 4 hrs)
- Cross-niche shadowing (first 5 calls per niche)
- Delivery (peaks after first close)
- Ops / planning / retro (typical 2 hrs)
Add the **shared cadence**: Mon planning meeting, Fri retro, first-5-calls observation rotation, first-proposal co-author rule, first-closed-deal kickoff.
Section 5 — Tool stack + cash budget
Build a `Required (week 1)` table with: tool name, monthly cost, owner, purpose. Include LinkedIn Sales Navigator (×2), Apollo, prospecting database (Crunchbase / Instrumentl / vertical-specific), Claude Pro/Team (×2), email warmup (Smartlead/Instantly), email verification (Hunter), Calendly Pro (×2), domain + landing pages.
Build a `One-time` table: insurance, conference travel buffer (max 1 trip), legal review, misc.
Build an `Approved-with-cause` table for tools that activate only on specific metric triggers (e.g., Google Ads after deal #2 closes; ahrefs after wk 8 if SEO is producing replies).
Subtotal monthly recurring + one-time = total cash for the horizon. Compare to the user's stated budget envelope. If over, recommend cuts.
Section 6 — Per-niche outbound systems
For each niche, produce a self-contained operational sub-section with:
6.x.1 ICP filter
Pull from the niche's `campaign-plan.md` and `market-research.md`. Express as a saved-search definition for the relevant prospecting tool (Crunchbase saved search, Instrumentl saved search, Apollo filter, etc.) — the actual filter syntax someone could paste into the tool.
Include manual disqualifiers (qualification rules the tool can't enforce).
State weekly output: ~N qualified leads/week.
6.x.2 Outbound sequence
Two or three concrete touches with **full email/LinkedIn templates**. Templates include:
- Subject line
- Body with merge tokens
- Trigger reference (the data point that makes this touch personalized)
- Wedge sentence (the differentiator from the niche's positioning)
- Clear ask
- Sender signature
Touch cadence: Day 0, Day +5, Day +10 (or per the niche's campaign-plan).
State stop rule: after touch N, no more outreach for 90 days.
State expected reply rate (from campaign-plan).
6.x.3 Discovery-call playbook
- Call length (30 or 45 min)
- Who drives, who observes (and the first-5-calls rotation)
- Section-by-section script outline with time allocations
- Qualification gate (4 must-have signals before moving to proposal)
- Failed-qualification soft close (long-term nurture motion)
6.x.4 Proposal + close playbook
- Proposal format (page count, format — Doc preferred over deck)
- Required sections
- Send-within timeline
- Closing terms (deposit %, NET payment, deliverable schedule, retainer minimums, termination notice)
- Auto-offer extensions (e.g., Type II retainer offered at Type I delivery)
6.x.5 Niche-specific channel mechanics
E.g., for a SOC2 niche: Vendor partner-program applications with specific URLs, expected timelines, follow-up cadence, and the binary kill-criterion week. For a grant-ops niche: Speaking + visibility submissions to GPA, AFP, Candid Learning. For other niches: whatever the campaign-plan flagged as the highest-leverage non-outbound channel.
6.x.6 Content cadence
- Anchor essay schedule (titles, weeks, distribution channels) — pull from campaign-plan
- LinkedIn organic posts (number per week, topic rotation, post times)
- Twitter/X if applicable
- Co-authoring rules across the founders
Section 7 — Cross-niche operational systems
7.1 CRM + pipeline tracking
Reference the existing Conversint CRM (`config/crm.json`, `lib/crm.py`, per AGENTS.md). State the pipeline stages, weekly snapshot routine, sheet URL surfacing rule.
7.2 Discovery-call observation rotation
State the first-5-calls-per-niche shadowing rule with explicit retro structure.
7.3 Weekly cadence
Hour-by-hour weekly schedule covering both founders' time blocks. Include the Mon planning meeting, daily outbound block, content publishing windows, Fri retro.
7.4 Metrics dashboard
Build a metrics table with current / wk-4 / wk-8 / wk-12 columns covering all the metrics from Section 1. State who updates it, where it lives (typically a shared Google Sheet), and the Friday-retro snapshot rule.
Section 8 — Week-by-week ops calendar
12 weeks (or chosen horizon) with day-by-day actions for week 1, then weekly summaries weeks 2–12.
For each week, list:
- Outbound touch volume target
- Content publication scheduled
- Niche-specific channel actions (partner-program follow-ups, audit-firm outreach, speaking submissions, etc.)
- Pipeline milestone targets (first proposal, first close, partner badge live, case study delivered)
- Metric review checkpoints (wk 4, wk 8, wk 12)
- Wk-8 binary gate(s) where applicable
End with the **wk-12 outcomes assessment**: decision meeting Saturday, decide continue / amend / activate Phase 2 / kill / pivot.
Section 9 — Delivery onboarding playbooks
For each niche, a step-by-step delivery motion for the first closed deal:
- Day 0 actions when deposit clears
- Kickoff call agenda
- Phase-by-phase deliverable schedule
- Auto-offer extensions at delivery completion
These exist so the cold-start motion doesn't break when the first close arrives and delivery hours suddenly compete with outbound.
Section 10 — Kill criteria + pivots
For each niche, list explicit kill criteria from the campaign-plan (typically "any 2 of 5 by week N"). Then a **pivot menu** with 2–3 alternatives:
- Pivot A: narrow the wedge
- Pivot B: change the channel mix
- Pivot C: kill the niche, reallocate hours
End with the **Phase-1 overall kill criterion**: at week 12, if outcomes-X and outcomes-Y both fail, full strategy review on Saturday.
Section 11 — Cash flow forecast
Week-by-week cash in / cash out / net / cumulative table for the full horizon. Use:
- Cash in: deposit timing per niche's terms (e.g., 50% on signing for sprints; setup fee + month-1 retainer for retainer niches)
- Cash out: tool subscriptions monthly + one-time costs in week 1
- Show a **target scenario** (per Section 1's "target" tier metrics) and explicitly state floor + stretch scenarios in a note below the table
These are target numbers, not guarantees. Be explicit.
Section 12 — Top 5 risks + mitigations
A 5-row table: risk / probability / impact / mitigation. Cover at minimum:
- Highest-leverage external dependency (partner badge, regulatory ruling, etc.)
- Quality / compliance error in delivery
- Outbound reply-rate floor
- First-deal delivery slip
- Founder burnout
Section 13 — The two ongoing decisions
13.1 Phase-2 niche activation trigger
State the explicit conditions (typically: 1 closed deal + 1 published case study + 1 referenceable client) and what activation looks like (typically: 25% of one founder's time on Phase-2 niche, ramping to 50% only after first discovery call).
13.2 Niche-pair revision trigger
State the wk-12 condition (e.g., "metrics missed by >50% on both niches → swap one for an unscreened candidate"). Reference the brainstormed-but-unscreened list in the triangle inventory.
Section 14 — Kickoff meeting agenda
60-min agenda with timeboxed sections:
| Time | Topic | Decision needed | |---|---|---| | 0–10 | Phase-1 commitment review | Confirm or amend | | 10–20 | Labor split | Confirm or amend | | 20–30 | Budget + tool stack | Approve | | 30–40 | Niche-specific decisions (per niche) | Lock | | 40–50 | Case-study target #1 selection (if applicable) | Lock with name + email | | 50–60 | Calendar holds | Set recurring blocks |
End with: after meeting, amend the playbook with any decisions, commit to main, send link to founders.
Appendix A — File index
Linked list of every committed deliverable: triangle inventory, all 3 files per niche.
Appendix B — Change log
Bullet list of dated revisions to this playbook.
Output rules
- **Format:** Markdown only. No DOCX, no date suffix in filename, no version suffix. Save at `research-thought-leadership/_inventory/<filename>.md`.
- **Voice:** Internal operations manual. Direct, specific, imperative. No client-facing voice. No marketing language.
- **Numbers:** Every dollar amount, time estimate, conversion rate, and reply-rate target must reconcile with the upstream campaign-plan. **Do not paper over math.** If the campaign-plan says "$18–32K CAC, 18–28 weeks," the kickoff playbook says the same — it does not soften it.
- **Templates:** Outbound email templates must be copy-pasteable, with merge tokens and trigger references. Do not write generic SaaS spam.
- **References:** Every section that pulls from an upstream file must link to the exact file path (relative path within the repo).
- **Honesty:** If a niche has weak triangle scoring (e.g., long sales cycle, high CAC), the kickoff playbook reflects it — both in the metrics targets and in the risks section.
Persistence
After writing the playbook:
- **Always commit to `main`** via the helper script — never leave the file untracked. Project rule per AGENTS.md: every markdown rewrite under `research-thought-leadership/` must be committed to main with `scripts/commit-md-to-main.sh`. The helper handles worktree-based commits, push, and surfaces a GitHub URL.
```bash scripts/commit-md-to-main.sh \ "research-thought-leadership/_inventory/<filename>.md" \ "Add cold-start kickoff playbook: <one-line scope>" ```
- **Surface the GitHub URL** the helper prints, as the shareable link.
- **Do not auto-invoke `/gdrive`.** Per AGENTS.md, document-generating skills must never auto-mirror to Drive. The user invokes Drive sync themselves if they want it.
- **Update the triangle inventory** if a Phase-1 commitment changes the active-deep-dive table or the cross-niche synthesis. Edit `research-thought-leadership/_inventory/triangle-comparison.md` and commit it in the same skill run.
Quality gate
Before returning, verify:
- [ ] Every metric in Section 1 reconciles with the upstream campaign-plan numbers
- [ ] Combined founder-time math in Section 2 is shown explicitly with the utilization percentage
- [ ] Section 4 labor split has hours-per-week breakdown for each founder summing to their weekly capacity
- [ ] Section 5 tool stack subtotal matches the cash-budget envelope (or explicitly flags overage)
- [ ] Section 6 outbound systems include full templates per niche, not placeholders
- [ ] Section 8 week-by-week calendar has day-level granularity for week 1
- [ ] Section 10 kill criteria match the upstream campaign-plan's kill criteria verbatim
- [ ] Section 11 cash flow uses the niches' actual deposit terms, not generic SaaS billing
- [ ] File is committed to `main` and the GitHub URL is surfaced
If any check fails, fix before declaring done.
Done state
Return to the user:
1. The GitHub URL from the helper script 2. A 5-bullet executive summary of what's in the doc: - Phase-1 niche pair (or trio) committed - Combined founder-hour utilization % - Wk-12 floor revenue target - The 3 decisions to lock at the kickoff meeting - Phase-2 activation trigger 3. The first 3 actions for week 1 (typically: subscribe to tools, submit partner applications, identify first case-study target) 4. **Optional next step:** offer to run `/gdrive` to mirror the doc to Drive if the user wants it, but do not auto-run it.