Smit Patel’s technical-product-gtm — 11 strategic Skills named after the failures, not the frameworks.

TL;DR

• 11 strategic Skills shipped as a Claude Code plugin via .claude-plugin/marketplace.json — version 3.0.0, MIT, zero tool dependencies. Pure prompt/framework skills.
• The failure-named heading discipline is the most copyable technique in the pack. “If Your MAP Hasn’t Been Updated in 3 Weeks, That Deal Is Dead” is a diagnostic; “Account Planning Framework” is a table of contents.
• The standout skill is ai-gtm (21KB, largest). The Copilot/Agent/Teammate positioning matrix maps each AI product category to a different buyer ceiling and a different price point.
• The most copyable specific tactic is Ceiling Moment Qualification — target users who’ve hit the incumbent’s limit. Patel claims 3 to 5x conversion lift; the mechanism is that ceiling pain is concrete and demonstrable, not latent.
• The pricing skill is the second strongest. Value Ratio formula: >10x = underpriced, 3-5x = healthy, <2x = wrong customer or wrong product. The “$15K → $45K, zero customers lost” anecdote is the structural argument.
• Where strategic-only skills hit a ceiling: zero tool dependencies means the skills stop at the doc layer. They don’t read your CRM, they don’t score your transcripts, they don’t run your campaigns. Decision support, not execution.
• What to borrow: the failure-named heading style applied to your own content, the Ceiling Moment Qualification language in your discovery scripts, the Value Ratio rubric in pricing reviews. What to leave: the assumption that publishing strategic content is the same as running the playbook.

Why read technical-product-gtm

Most published GTM content sits in one of two failure modes. Either it is too high-level — frameworks without failure modes, decision trees without decisions, decks without the war underneath — or it is too narrow — war stories without abstraction, a single deal recounted as if it generalizes. Patel’s skills sit in the middle. The voice (“Security approved, but operations rejected us”; “Press ≠ Growth”; “$15K → $45K, zero customers lost”) is the rare combination of specific failure mode and generalizable lesson, and that combination is what makes the pack worth a read even if you never invoke the skills directly from Claude Code.

The structural lesson generalizes across content domains and is, in our view, worth more than any single skill in the pack. We’ll get to it. First, the survey: 11 skills, each ~10 to 20KB, each following the same template — Core Frameworks → Decision Trees → Common Mistakes → Quick Reference → Related Skills. Every skill ships with a YAML frontmatter, MIT license, and a single SKILL.md as the entire payload. There are no MCP servers, no API calls, no CRM integrations. The plugin is a directory of strategic essays addressed to the operator who is about to make a decision, and the strategic depth across the 11 essays is unusually consistent.

One upstream point that matters for the rest of this chapter. Patel writes from inside the operator’s chair. The voice is not the consultant’s voice (“a four-phase framework for cross-functional alignment”) and it is not the academic’s voice (“recent work in pricing theory suggests”). It is the voice of someone who has shipped, missed, recalibrated, and shipped again. That voice is the entire reason the failure-named heading discipline lands — without it, the headings read as performative. With it, they read as scar tissue.

The failure-named heading discipline

This is the most important section of the article and the single technique worth stealing from the repo. Patel’s headings are named after the failure pattern the operator is about to encounter, not the framework that resolves it. A small sample, quoted verbatim from across the SKILL.md files:

• “If Your MAP Hasn’t Been Updated in 3 Weeks, That Deal Is Dead” (skills/enterprise-account-planning/SKILL.md)
• “Press ≠ Growth (The Launch That Got 12 Signups)” (skills/0-to-1-launch/SKILL.md)
• “The Week 4 Ghosting Problem” (skills/enterprise-onboarding/SKILL.md)
• “The Week 12 Adoption Cliff” (skills/enterprise-onboarding/SKILL.md)
• “Security approved, but operations rejected us” (skills/enterprise-account-planning/SKILL.md)
• “The $5K → $50K Inflection” (skills/product-led-growth/SKILL.md)

The mechanism is a property of human pattern recognition. Operators do not store knowledge as “Account Planning Framework” or “Onboarding Methodology.” They store it as “the last three deals that died at MAP staleness” or “the customer who ghosted in week four and never came back.” The failure-named heading anchors on the storage format the operator’s memory is already using, and the framework underneath the heading slots in as the resolution to a pattern the operator has already pattern-matched on.

A reader who sees “If Your MAP Hasn’t Been Updated in 3 Weeks, That Deal Is Dead” immediately scans their own pipeline for stale MAPs. The cost of skimming the section is zero because the question “do I have any stale MAPs” is so concrete that even a no-answer is informational. A reader who sees “Account Planning Framework” skims to the bullet points, recognizes the standard MEDDICC structure, and closes the tab. The first heading is a diagnostic. The second is a table of contents. Diagnostics earn read-through; tables of contents earn bookmarks that never get opened.

The discipline has a second property that matters for skill-pack authors specifically. A Claude Code skill is invoked by a model that matches against the skill description and the section headings. A heading that names a failure mode the user is describing (“my MAP hasn’t been updated in a few weeks and I’m worried about the deal”) has a substantially better match against the user’s utterance than a heading that names the framework (“Account Planning Framework”). The discipline is a retrieval optimization as well as a reader optimization. Patel almost certainly did not write it for that reason, but the property holds either way.

The discipline is also constrained. A heading like “The Week 4 Ghosting Problem” is empty if you do not have data on Week 4 ghosting. Copy the discipline; do not copy the specific failures. The headings have to be named after failures the author has actually observed and the reader can actually recognize. A heading invented to sound diagnostic — “The Tuesday Morning Inflection Point” — reads as performative and damages the rest of the pack. Patel’s headings work because the failures behind them are real enough to be specific. The discipline is not stylistic. It is a function of the depth of the underlying experience.

The standout skill: ai-gtm

The largest skill in the pack (21KB) and the one most worth reading even if you adopt none of the others. skills/ai-gtm/SKILL.md centers on a positioning matrix — Copilot, Agent, Teammate — that maps each AI product category to a different buyer ceiling, a different price point, and a different accountability model.

A Copilot is a suggestion engine. The buyer is the individual user; the accountability stays with the user; the pricing is per-seat in the $20 to $100 per month range. The buyer ceiling is the moment the user wants the system to act, not suggest. A Copilot positioning above its ceiling reads as overpromising; a Copilot positioned at its ceiling reads as honest and prices accordingly.

An Agent is an action engine. The buyer is the team lead; the accountability is partially delegated to the system; the pricing is per-workflow or per-outcome in the $500 to $5K per month range. The buyer ceiling is the moment the agent needs to operate without supervision across multiple workflows. An Agent positioning above its ceiling reads as a Teammate that hasn’t earned the trust; an Agent positioned at its ceiling reads as a force multiplier on a defined function.

A Teammate is a fully delegated worker. The buyer is the executive who would otherwise hire a person; the accountability is fully delegated; the pricing is per-headcount-equivalent in the $5K to $50K per month range. The buyer ceiling is whatever a competent human would do in the same role. A Teammate positioning has to clear the “would I trust this with the actual job” bar, and the pricing reflects the savings against the alternative hire, not the cost of the underlying tokens.

The non-obvious move is that each category maps to a different accountability cascade. The Copilot vendor is accountable for suggestion quality. The Agent vendor is accountable for workflow completion. The Teammate vendor is accountable for the outcome a human would have produced. Operators who position their product one tier above its actual accountability cascade collapse under the first failure — the customer ascribes Teammate accountability to a product that ships at Agent accountability, the failure is read as a fundamental product flaw, and the contract churns. Operators who position at or one tier below the actual cascade build trust on the upside.

The single most copyable tactic in the entire repo also lives in this skill: Ceiling Moment Qualification. The premise is that a buyer who has hit the incumbent’s limit converts at 3 to 5x the rate of a buyer who is exploring without a ceiling. The mechanism is concrete: ceiling pain is demonstrable (the buyer can show you the exact workflow that broke); latent need is not (the buyer can describe the problem but cannot prove it costs anything). A discovery script tuned to surface ceiling moments — “tell me about the last time the current tool couldn’t do what you needed it to do” — separates buyers who will close from buyers who will explore. We’ve added the language to our own discovery scripts; the empirical lift on the calls we’ve audited is consistent with Patel’s claim.

Ceiling Moment Qualification has a structural constraint: it requires a clear incumbent. In a green-field market — no incumbent, no ceiling to hit — the tactic does not apply. Operators who copy the language into a green-field discovery script produce flat results because there is no ceiling for the buyer to describe. The tactic is a function of the market structure, not a universal disqualifier.

The pricing skill

skills/technical-product-pricing/SKILL.mdis the second strongest skill in the pack. The core construct is the Value Ratio — the ratio of the buyer’s captured value to the price. >10x = underpriced (you are leaving money on the table and signaling commodity status); 3 to 5x = healthy (the buyer has clear ROI and the vendor captures a meaningful share); <2x = wrong customer or wrong product (the buyer either does not capture enough value to justify the price or has been mis-targeted into a tier that does not fit).

The structural argument behind the formula is that price is a positioning signal as much as a revenue capture mechanism. An operator who prices at a 20x Value Ratio is signaling that the product is a commodity — the buyer infers that the vendor does not believe the product captures enough value to justify a higher price, and the buyer’s perception of the product compresses to match. An operator who prices at a 3 to 5x Value Ratio is signaling that the vendor understands what the buyer is buying and has calibrated the price to the buyer’s economics. The signal is read pre-purchase, and it changes the buyer’s evaluation criteria.

The anecdote that anchors the skill is “$15K → $45K, zero customers lost.” The structural claim is that a 3x price increase on a properly-calibrated Value Ratio produced no customer loss because the original price was below the Value Ratio band — the buyers were already capturing far more value than they were paying for, and the price increase moved the vendor from underpriced to healthy without crossing into wrong-customer territory. Anonymized, unverifiable, but structurally consistent with the formula. The operator who reads the anecdote as a recipe (“triple my price and no one will leave”) is reading it wrong. The operator who reads it as calibration (“if my Value Ratio is 20x, a 2 to 3x price increase should not produce churn”) is reading it right.

The Value Ratio is most useful in customer audits. The question we ask during pricing reviews is whether the customer’s current price sits in the 3 to 5x band, the >10x band, or the <2x band, and the answer determines what move to make. Customers in the >10x band are leaving money on the table and we recommend a structured price increase against the existing book. Customers in the <2x band are mis-targeted and the conversation is about either ICP refinement or product repositioning, not pricing. The rubric is calibration, not contract pricing.

The 11 skills in survey form

A one-paragraph orientation to each skill. The two we’ve already covered — ai-gtm and technical-product-pricing — are the strongest. enterprise-account-planning and enterprise-onboarding are the most operationally specific. The rest are useful at varying degrees of abstraction.

• 0-to-1-launch.The thesis is that direct outreach beats press for early-stage launches and the launch-as-an-event is a category error. The anchor anecdote is “12 signups from 50K impressions” — a press hit produced volume without conversion because the audience was not pre-qualified. The skill prescribes 2-week experiment cycles in place of single-launch moments. The failure-named heading is “Press ≠ Growth.”
• ai-gtm. Covered above. Copilot/Agent/Teammate positioning matrix, accountability cascade, AI trust ladder, Ceiling Moment Qualification. The single largest skill (21KB) and the one most worth reading. skills/ai-gtm/SKILL.md.
• board-and-investor-communication.The “Three Things” narrative structure — every board update reduces to three things, and the discipline is in the reduction. A 4-tier metric hierarchy separates north-star from input metrics. The standout prescription: pre-brief bad news to individual board members 48 hours before the meeting, so the room reaction is calibrated rather than reactive.
• developer-ecosystem. Open vs curated marketplace tradeoffs, a 3-year arc for university/student programs, and a 5-stage developer journey from awareness to advocacy. The skill is strongest on the marketplace tradeoff — open marketplaces produce volume and noise; curated marketplaces produce quality and gatekeeping risk. The choice is a function of the developer surface area, not a universal answer.
• enterprise-account-planning. MEDDICC-anchored but operationally specific. The standout heading is “If Your MAP Hasn’t Been Updated in 3 Weeks, That Deal Is Dead” — the failure-named diagnostic that turns MAP staleness from a process gap into a deal signal. Personal Win Mapping treats each champion as a person with a career trajectory, not a logo. skills/enterprise-account-planning/SKILL.md.
• enterprise-onboarding. A 4-phase onboarding flow anchored on two failure patterns — the Week 4 Ghosting Problem (the customer stops responding to status emails around week four) and the Week 12 Adoption Cliff (active usage falls off a cliff around the 90-day mark). Both are diagnostic checkpoints. The skill is the most operationally specific in the pack.
• operating-cadence. A 5-level meeting architecture from daily standups to quarterly board reviews, with decision velocity and role clarity as the two metrics that diagnose cadence dysfunction. The skill is strongest on the decision velocity construct — a team whose meetings produce decisions is healthy; a team whose meetings produce status updates is decaying.
• partnership-architecture. A 3-tier partnership model (technology, channel, strategic) with a skin-in-the-game test as the disqualifier. The thesis: partnerships without revenue commitments are press releases. The crawl-walk-run sequence is the standard staged commitment escalator and lands cleanly because the skin-in-the-game test gates the escalation.
• positioning-strategy.Centers on “the word change” — the observation that repositioning is often a single-word substitution that changes the buyer’s mental model (Autonomous → Teammate is the worked example). The 3-layer positioning architecture (category, point of view, proof) is the standard frame; the word-change construct is the move worth borrowing.
• product-led-growth.Anchored on the “$5K → $50K inflection” — the price band at which PLG self-serve breaks down and a sales-assist motion becomes necessary. The growth equation decomposes PLG conversion into activation, retention, and expansion. The standout anecdote is a parallel test in which sales-led won 10x against pure PLG on a product the operator had assumed was PLG-native.
• technical-product-pricing. Covered above. Value Ratio formula, the “$15K → $45K, zero customers lost” anchor, the price-as-positioning-signal argument. skills/technical-product-pricing/SKILL.md.

Where technical-product-gtm stops short

Five honest gaps an operator hits the moment they try to run the pack as a system instead of a reading list:

• Zero tool dependencies. Pure prompt/framework skills are portable but stop at the doc layer. They do not read your CRM, they do not score your transcripts, they do not run your campaigns. The skills produce decisions; the execution layer is yours. This is both a strength (no integration debt, no vendor lock-in) and the limit. A founder who invokes ai-gtm from Claude Code gets a calibrated decision on how to position their product. The next move — repricing the page, rewriting the deck, retraining the sellers, sequencing the rollout — is unsupported.
• Provenance gated to README.The Datadog and Postman references appear in the repo’s README and AGENTS.md as credential context for the author. They do not appear in any SKILL.md. The operator who reads only the skills sees anonymized anecdotes (“a platform company that raised enterprise tier from $15K to $45K”) without the ability to audit the source. The provenance is credible but unverifiable from inside the skill.
• Anonymized anecdotes.The “$15K → $45K, zero customers lost” story is structurally true but operationally unverifiable. Operators who adopt the pattern without the underlying data risk applying it to a context where the math does not hold — a buyer base that is closer to the <2x band than the >10x band, where a 3x price increase produces 30% churn instead of zero. The anecdotes are calibration. They are not contract pricing.
• No closed feedback loop. The skills are static documents. They do not update from how your team actually performs. The Value Ratio you calibrate to today does not refresh from the actual win/loss data you accumulate next quarter. The Ceiling Moment Qualification language you adopt in your discovery scripts does not auto-tune from the calls that closed. The pack is a snapshot; the loop is yours to build.
• Strategic abstraction.The skills are at the right level for a founder making a decision and the wrong level for a rep on a Tuesday call. A rep does not need “the Copilot/Agent/Teammate positioning matrix.” A rep needs the seven-sentence pitch in the founder’s voice that operationalizes the matrix on a discovery call. The translation from strategic skill to executional artifact is the work the operator has to do, and the pack does not do it for them.

None of these are criticisms of the pack. Patel scoped at the strategic-decision layer — that is the layer where the edge sits. The gaps are the execution layer underneath, and the operator who closes them is the operator who converts strategic skills into shipped campaigns.

How Allston applies it

The failure-named heading discipline runs across our own library. We audited the 86 guides in the library against the discipline — diagnostic heading or table-of-contents heading — and retitled the 15 to 20% that read as the latter. The dwell-time lift on the retitled guides was measurable inside two weeks. The discipline is the single most generalizable lesson in Patel’s pack and the one we have spent the most internal time operationalizing.

The Ceiling Moment Qualification tactic is built into our discovery scripts as a specific cue. The prompt — “tell me about the last time the current tool couldn’t do what you needed it to do” — runs on every first-touch discovery call across customer engagements, and the answer is scored as a discrete dimension on the call. Buyers who answer with a concrete ceiling moment convert at the rate Patel describes. Buyers who cannot name a ceiling moment are routed back to nurture rather than to a follow-up call.

The Value Ratio is the rubric we use during customer audits when the question is whether the customer is underpriced. We do not use it as a contract-pricing tool — the math is too dependent on the buyer’s captured-value estimate to be precise — but we use it as a calibration check. A customer whose Value Ratio sits at 15x is leaving money on the table, and the audit recommends a staged price increase. A customer whose Value Ratio sits at 1.5x is mis-targeted, and the audit recommends ICP refinement before any pricing move.

The skills themselves are decision support. The execution layer — running the campaigns that produce the meetings, scoring the calls that surface the ceiling moments, routing the replies, repricing the book — is the work we do across customer engagements. The split is clean: read Patel for the decisions, run the operations to convert the decisions into revenue. We do not duplicate Patel’s strategic depth, and Patel does not duplicate our operational layer. The two compose.

Operator failures when adopting technical-product-gtm patterns

• Reading the skills as prescriptions rather than diagnostics. Patel writes patterns; operators want recipes. The Copilot/Agent/Teammate matrix is a calibration tool — it tells you which category your product sits in and what that implies for accountability and pricing. It does not tell you how to position your specific product against your specific competitor. The patterns are the input; the specific positioning move is yours.
• Adopting the failure-named heading style without the underlying failure data.A heading like “The Week 4 Ghosting Problem” is empty if you do not have data on Week 4 ghosting. Copying the discipline produces strong content; copying the specific failures without the underlying experience produces hollow content that the reader detects within two sentences. Steal the discipline. Earn your own failures.
• Treating the Value Ratio as exact rather than directional. The math is calibration, not contract pricing. An operator who triples a price because the formula said the current ratio was 15x — without auditing whether the buyer captures value in the way the formula assumes — produces churn the formula did not predict. The ratio is a sanity check on a pricing direction, not the pricing decision.
• Applying Ceiling Moment Qualification to a market without a clear incumbent. The tactic requires a ceiling. In a green-field market the buyer cannot describe a moment when the current tool failed, because there is no current tool. Operators who apply the language in green-field discovery produce flat answers and conclude the tactic does not work, when the actual conclusion is that the market structure does not support it.
• Using the strategic skills as content rather than as decision support. Publishing the Copilot/Agent/Teammate framework on your blog is not the same as running it as a checklist before a positioning meeting. The skills earn their keep when they are invoked at the moment of decision, not when they are repurposed as marketing. Operators who treat strategic content as a publishing strategy get attention; operators who treat it as a decision rubric get calibration.
• Skipping the operational layer.Strategic skills produce decisions. Execution produces revenue. An operator who reads the pack, calibrates a Value Ratio, and stops there has not changed the book’s revenue. The price has to be moved on the page, the email sequence has to be re-sent to the existing accounts, the renewal conversations have to be scripted. The strategic decision is one of ten artifacts the operator has to produce, and the other nine are the work.

Adoption checklist

• The failure-named heading discipline has been applied to your own content — at least one audit of an existing guide has been completed and the table-of-contents headings have been retitled as diagnostics
• The Copilot/Agent/Teammate positioning matrix has been run against your product, the current tier has been named explicitly, and the accountability cascade has been audited against the current contract language
• Ceiling Moment Qualification language has been added to discovery scripts and the answer is scored as a discrete dimension on every first-touch call
• The Value Ratio has been calculated for the current pricing across at least three representative customer segments, and the band (>10x, 3-5x, <2x) is known for each
• A MAP staleness audit has been run on the active pipeline — deals with MAPs not updated in 3 weeks have been triaged and either reactivated or closed-lost
• A Week 4 onboarding check has been instrumented in the customer-success motion — customers who have not responded by week four trigger a specific escalation, not a status email
• The Week 12 adoption cliff has been instrumented as a leading indicator on the customer health score
• The “Three Things” narrative discipline has been applied to the next board update — the update has been reduced to three things before the slides are built
• The pre-brief discipline on bad news has been adopted — board members are briefed individually 48 hours before any meeting that contains bad news
• The strategic skills have been moved from a reading list to a decision rubric — each skill has been bookmarked against a specific recurring decision (positioning review, pricing review, onboarding postmortem) and is invoked at that cadence

Where this fits

technical-product-gtm is one chapter in our framework analysis series, and it sits at the strategic-decision layer that the operational frameworks build off. Rob Snyder’s PULL framework is the discovery-layer methodology that surfaces the Ceiling Moments — PULL’s four-question sequence is the operational scaffold that Patel’s Ceiling Moment Qualification language slots into. SalesGraphis the upstream research scaffold that produces the named-account view against which Patel’s account-planning skill runs.

On the in-house library specifically, our principles of cold copychapter uses the same failure-named heading style — “AI-slop failure modes,” “What kills cold emails,” “The voice problem” — that Patel uses across the skill pack. The discipline is the same; the domain is different. The pair is the input set for an operator who wants the strategic decision layer (Patel) and the executional copy layer (our library) operating off the same heading discipline.

Related chapters

• Rob Snyder’s PULL framework — the discovery-layer methodology that surfaces Ceiling Moments on the call.
• SalesGraph — the upstream research scaffold against which strategic account planning runs.
• Richard Makara’s GTM Context OS — the operational scaffold that converts strategic skills into a working context system.
• The principles of cold copy — the in-house library chapter that operationalizes the failure-named heading discipline at the copy layer.