cmposer

name	cmposer
description	Detect early customers and design partners that will burn you: tire-kickers with no real pain, champions with no budget authority, roadmap hijackers, free-pilot-forever logos, vanity names that never convert, and counterparties who squeeze or ghost early-stage vendors. Usage: /cmposer <Account Name>. Runs 8 evidence-based checks, scores a Fit Score (0–100), and recommends whether to pursue. Saves to cmposer-<account>.md.
license	MIT
compatibility	Requires Claude Code with WebSearch and WebFetch. Internet connection required.
allowed-tools	WebSearch WebFetch Read Write
metadata	{"author":"3Flux","version":"1.0","workflow-step":"3 — run before /cmmatch to screen out bad-fit accounts"}

/cmposer — Bad-Fit Account Detector

Run a structured 8-check screen against a candidate design partner or early customer to determine whether the relationship will advance the startup or quietly bleed it. Produces a Fit Score and a verdict before a founder pours weeks of meetings, custom work, and roadmap concessions into an account that was never going to buy — or worse, one that will distort the product for everyone else.

The wrong early customer is a slow-motion disaster: the pilot that never defines "yes," the champion who can't sign, the whale whose one-off demands become your backlog, the famous logo that proves nothing because no one there actually uses the product. This screens for all of it.

Invocation

/cmposer <Account Name>

Examples: /cmposer "Acme Utilities" or /cmposer Stripe

Execution Steps

Step 1 — Load Startup Context

Read STARTUP_PROFILE.md (and cmstrat.md if present). Extract and hold in context:

• The beachhead segment and Ideal Design Partner Profile
• The specific problem and the pain it relieves
• The buyer role, likely champion, and price point
• Runway / how long a sales cycle the startup can survive
• Whether the startup is selling a paid pilot, a design-partner agreement, or a contract
• Reference-value needs (which logos matter for the next wave)

This context calibrates the checks. A 9-month procurement cycle is fatal for a startup with 8 months of runway but fine for one that's well-funded. A vanity logo is worthless to a startup that needs revenue but valuable to one that needs signal for a raise. Flag relative risk, not just absolute signals.

If STARTUP_PROFILE.md does not exist, stop and tell the user to create it first.

Derive the account slug and output filename: lowercase, strip spaces and punctuation.

• Stripe → cmposer-stripe.md
• Acme Utilities → cmposer-acme-utilities.md

Step 2 — Gather Raw Intel

Use WebSearch and WebFetch across exactly these 6 angles. Run all 6 before scoring.

1. "<Account>" "<problem area>" initiative priority strategy 2025 2026 — is the pain real and active?
2. "<Account>" budget spend procurement "<category>" — budget reality and how they buy
3. "<Account>" pilot vendor startup partnership case study — do they actually adopt/convert, or pilot-and-stall?
4. "<Account>" "<buyer role>" leadership champion LinkedIn — who decides and who could champion
5. "<Account>" vendor relationship payment terms lawsuit dispute layoffs — counterparty risk and stability
6. site:<account-website> OR news "<Account>" "<problem area>" — confirm acuity and timing

Pull named people, budget signals, dates, and adoption history wherever possible.

If fewer than 3 of the 6 angles return usable data (account is small, private, or opaque): note this explicitly in the report header and default the verdict to Caution. Do not score what you cannot evidence — but also note that thin public data is itself a mild reachability signal.

Step 3 — Run the 8 Fit Checks

Run each check in sequence. For each, produce:

• Signal: 🟢 Green / 🟡 Yellow / 🔴 Red
• Finding: 1–2 sentences with specifics — names, dates, budget/initiative evidence
• Evidence: the source (URL, quote, or named publication)

---

Check 1: Pain Acuity (20 pts)

What it tests: Is the problem you solve a top-priority, actively-funded pain for them — or a nice-to-have they'll nod at and never prioritize? Low pain is the #1 cause of stalled early deals.

• 🟢 Evidence the problem is a current, named priority (public initiative, hiring, exec statement, regulatory pressure)
• 🟡 Plausible pain but no active signal — they have the problem but may not be prioritizing it
• 🔴 No evidence of acute pain, or the problem is clearly peripheral to their actual priorities

Points: 🟢 = 20 · 🟡 = 10 · 🔴 = 0

---

Check 2: Budget & Authority (20 pts)

What it tests: Is there a real budget line, and a champion senior enough to commit it? "We love it, let me check with..." forever is the sound of no authority.

• 🟢 Identifiable budget for this category and a reachable champion with signing authority
• 🟡 Budget exists somewhere but authority is unclear, or the likely champion needs sign-off you can't see
• 🔴 No evidence of budget, or the only contact is too junior to commit — high risk of an endless cycle

Points: 🟢 = 20 · 🟡 = 10 · 🔴 = 0

---

Check 3: Decision Velocity (10 pts)

What it tests: Can they decide and contract inside a timeframe the startup's runway survives? Enterprise procurement and legal can outlast a seed round.

• 🟢 Evidence of fast vendor adoption or a lightweight buying process — they can move in weeks
• 🟡 Moderate process — quarters, not weeks; survivable but plan for it
• 🔴 Heavyweight procurement/legal/security review likely to exceed the startup's runway — a velocity trap

Points: 🟢 = 10 · 🟡 = 5 · 🔴 = 0

---

Check 4: Roadmap Alignment (15 pts)

What it tests: Does what this account needs match where the product is going — or will they pull you into one-off custom work that serves no one else? The roadmap hijacker builds you a product for a market of one.

• 🟢 Their needs map onto the planned roadmap and would benefit the broader ICP
• 🟡 Mostly aligned with some bespoke asks — manageable if scoped and time-boxed
• 🔴 Their core need requires custom work off the roadmap — a distraction that distorts the product

Points: 🟢 = 15 · 🟡 = 8 · 🔴 = 0

---

Check 5: Conversion Intent (15 pts)

What it tests: Will a pilot actually become paid? Look for a culture of buying tools vs. endless free evaluations. Pilot-to-paid history is the best predictor.

• 🟢 Evidence they convert pilots to paid contracts and pay vendors fairly
• 🟡 Mixed or unknown — they pilot, but conversion history is unclear
• 🔴 Pattern of perpetual free pilots, "innovation theater," or extracting vendor work without buying — a free-forever trap

Points: 🟢 = 15 · 🟡 = 8 · 🔴 = 0

---

Check 6: Reference Value (10 pts)

What it tests: If you win them, does it move the next ten buyers? A logo only has value if it converts to a usable reference, quote, or intro — not just a badge.

• 🟢 Respected logo in the beachhead and plausibly willing to be a real, usable reference
• 🟡 Decent reference value but uncertain whether they'd go public, or limited reach in your segment
• 🔴 Vanity logo — impressive name but off-segment, or known never to provide references — proves nothing

Points: 🟢 = 10 · 🟡 = 5 · 🔴 = 0

---

Check 7: Engagement Capacity (5 pts)

What it tests: For a design partner specifically — will they actually give time, candid feedback, and data access? A design partner who signs and goes dark gives you a logo and no learning.

• 🟢 Evidence of an engaged, hands-on team with capacity to co-build (innovation team, prior partnerships)
• 🟡 Willing but stretched — engagement may be inconsistent
• 🔴 No capacity or culture for hands-on partnership — likely to sign and disappear

Points: 🟢 = 5 · 🟡 = 3 · 🔴 = 0

---

Check 8: Counterparty Health (5 pts)

What it tests: Are they a fair, stable counterparty? Search for a history of squeezing small vendors, abusive procurement, payment disputes, instability, or layoffs that could kill the deal mid-flight.

• 🟢 Stable, fair reputation with vendors; no red flags
• 🟡 Thin or mixed signal — nothing damning, but unverified
• 🔴 Documented pattern of squeezing/ghosting vendors, payment disputes, or instability threatening the deal

Points: 🟢 = 5 · 🟡 = 3 · 🔴 = 0

---

Step 4 — Score and Verdict

Sum the points across all 8 checks.

Fit Score: XX/100 (higher = better fit, lower bad-fit risk)

Show the per-check breakdown in a table:

Check	Signal	Points Earned / Max
1. Pain Acuity	🟢/🟡/🔴	X / 20
2. Budget & Authority	🟢/🟡/🔴	X / 20
3. Decision Velocity	🟢/🟡/🔴	X / 10
4. Roadmap Alignment	🟢/🟡/🔴	X / 15
5. Conversion Intent	🟢/🟡/🔴	X / 15
6. Reference Value	🟢/🟡/🔴	X / 10
7. Engagement Capacity	🟢/🟡/🔴	X / 5
8. Counterparty Health	🟢/🟡/🔴	X / 5
Total		XX / 100

Verdict:

Score	Verdict	What it means
80–100	Strong Fit	Real pain, real budget, real reference value — pursue now
60–79	Workable	Yellow flags present; pursue if a champion path exists, mitigate the gaps
40–59	Caution	Material bad-fit signals; only pursue with a warm path and eyes open
0–39	Walk Away	The relationship will cost more than it returns — redirect the energy

Step 5 — Save and Report

Write the full check report to cmposer-<account>.md.

Then tell the user:

• Strong Fit (80–100): "This account is worth real investment. Run /cmmatch <account> to map the deal and the champion."
• Workable (60–79): "Pursue, but mitigate the flagged gaps first. Run /cmmatch <account> to plan around them."
• Caution (40–59): "Only pursue with a warm path and a time-boxed pilot with a written success definition. Otherwise redirect."
• Walk Away (0–39): "Skip it — the flags (e.g., [top red flag]) will burn runway. Run /cmlist for better-fit accounts."

---

Output Format

When producing the output file, read the exact template from references/output-format.md.

---

Output Rules

• Every check must cite named evidence — URLs, dates, named people, budget/initiative signals. If absent, say what was searched and why it came up empty.
• Show the per-check score table before the check details.
• If an account has <3 angles with usable data, print this warning at the top: ⚠️ Limited public data — fewer than 3 checks could be fully evidenced. Verdict defaulted to Caution regardless of score.
• The Fit Score measures bad-fit risk — it is not a measure of how much you want them. A dream logo can be a Walk Away.
• Total word budget for check details: ≤700 words. The score table and verdict rationale are not counted.
• Do not recommend /cmmatch until the verdict is Workable or Strong Fit.
• Name the single biggest red flag explicitly in the Walk Away / Caution recommendation.