startups

Plugin startups from marketplace official-business-plugins in web-lifter/official-business-plugins. Commands read from the repo's manifest — 8 days ago.

Skills (69)

• adr-writer

  Write an Architecture Decision Record using the standard template — context, decision, consequences, alternatives, status. Used by every other planning skill that lands a decision. Writes 09-mvp/architecture/ADR-NNN-<slug>.md.

• api-design

  REST and/or GraphQL API design — route inventory, OpenAPI schema (or GraphQL schema), error contracts, auth and rate-limit decisions, idempotency rules. Writes 09-mvp/architecture/api-design.md.

• architecture-design

  Produce a system-architecture document — Mermaid component diagram, data-flow diagram, integration surface, sync/async boundaries. One file per architectural decision under 09-mvp/architecture/.

• auth-model-design

  Design the auth.users / profiles / roles / RLS-policy stack for typical SaaS patterns. Output is SQL plus a policy table. Supabase MCP optional. Read-only.

• bmc-build

  Construct the initial Business Model Canvas from venture vision and customer-discovery state. Tags every cell as hypothesis or fact. Writes 05-business-model/bmc-vN.md (or 01-hypotheses/bmc-v1.md).

• bmc-front-back-split

  Render the latest BMC into front-stage (right side — what the customer sees) and back-stage (left side — what the business does) views, so users can sequence work appropriately. Pure derivation — never modifies the source BMC.

• bmc-link-vpc

  Link each VPC to the segment + value-propositions cell of the latest BMC. Adds bidirectional links so the VPC and BMC stay in sync. Read-mostly — only adds links; never replaces cell content.

• bmc-revenue-cost-sketch

  Back-of-envelope pass at the BMC's Revenue Streams and Cost Structure using customer profile, channels, and pricing hypotheses. A sketch, not a financial model. Delegates to economics:unit-economics-calculator for the rigorous version.

• bmc-update

  Apply changes from the hypothesis register to the BMC — when a hypothesis flips to accepted/refuted, update the relevant cell and bump to v(N+1). Tracks the diff in a changelog. Marks the prior version superseded.

• channel-select

  Match the venture to direct vs indirect channels. Outputs a channel strategy ranking primary, secondary, and tertiary channels with rationale. Writes 06-relationships-channels/channel-strategy.md.

• churn-model

  Model customer survival and expected lifetime from a stated retention rate; distinguish assumptions from observed cohorts and compute gross-margin-adjusted economics.

• cloudflare-deploy-plan

  Cloudflare deployment plan — Workers vs Pages, D1 vs Hyperdrive, KV vs R2, queues, bindings, wrangler.toml. Optionally probes the user's account via the Cloudflare MCP (read-only). Mutations would require the connector-confirmation flow.

• competitor-bmc-shadow

  Produce a shadow Business Model Canvas for a top competitor — what we infer about their nine cells. Writes 05-business-model/shadow-<competitor>-vN.md. Every cell tagged inferred.

• competitor-discover

  Surface competitors across five sources — direct, indirect, substitutes, current workaround, do-nothing. Web-search required; cite every competitor with a URL.

• competitor-insights

  Synthesise across competitor-table, all SWOTs, and shadow BMCs — what did we learn, what does that change in our hypotheses? Outputs 04-competitors/insights.md plus proposed register updates.

• competitor-table-build

  Populate the canonical competitor table — name, category, segments served, pricing, features, source URL, verified date. One table per venture.

• converge-ideas

  Take the diverged set and converge to 1–3 finalist concepts. Applies four filters (segment fit, hypothesis test value, build cost, novelty wedge). Outputs 08-prototype/converged-<date>.md plus proposed test cards for the finalists.

• customer-discovery-status

  Run the four-question customer-discovery gate and emit a RAG-style readiness report. Blocking — MVP-planning skills check this gate before running. Overrides log automatically to log.md.

• customer-profile-build

  Build the right half of the Value Proposition Canvas for one segment — customer jobs, pains, gains. Each item tagged functional/social/emotional and prioritised high/medium/low. Updates segments/<slug>/profile.md.

• customer-segment-define

  Walk the user through specifying customer segments — occupation, demographics, problems they have. Distinguishes users from paying customers per Blank's customer-development model. Scaffolds the segment folder with required files. Optional richer personas via marketing.

• data-model-from-vpc

  Translate the value-map outputs and customer-profile outputs into a draft entity-relationship model. Bridges value-proposition → schema work. Output is Mermaid ERD plus an entity list with attributes, types, and relationships. No connector calls.

• digital-prototype

  Drive the digital prototype workflow — delegates to figma-design-handoff for Figma metadata and tokens, captures screenshots, produces a feedback-collection plan. Writes 08-prototype/digital/<slug>/README.md.

• divergent-ideate

  Run a structured divergence pass — generate many ideas without critique. Constraints kept light, judgement deferred. Outputs a ranked-by-novelty list under 08-prototype/divergent-<date>.md.

• early-adopter-profile

  Identify the early-adopter (earlyvangelist) sub-set of a segment using Blank's five criteria — has the problem, knows it, looking for a solution, cobbled a workaround, has budget. Writes segments/<slug>/early-adopters.md.

• experiment-data-collection-plan

  Given a test card, design the data capture — source, table/event, query for "are we right?" Pairs with experimentation/test-card-build to make the experiment instrumentable.

• experiment-design

  Pick the best experiment type for a hypothesis — interview, survey, smoke test, concierge, fake door, Wizard of Oz, A/B, pre-order, LOI. Recommends with rationale and drafts for /test-card-build.

• experiment-prioritise

  Rank open hypotheses by risk × impact × ease-of-test with the formula explicit. Outputs a ranked list of what to test next. Read-only.

• experiment-run-tracker

  Log experiment execution against an open test card — dates, who was contacted, raw data observations, current status. Distinct from interview-log (raw notes per session); this tracks the experiment as a unit. Append-only.

• figma-design-handoff

  Pull metadata from a Figma file via the Figma MCP — file metadata, design context, variables, screenshots, libraries. Produces a component spec, proposes a code-connect map, surfaces design tokens. Read-only with graceful degrade.

• funnel-instrumentation-spec

  Produce a concrete event/property spec from a funnel-model output. Translates Awareness → Sign-up → Activation → Conversion → Retention into named events with properties. Pairs with mvp-analytics-plan.

• funnel-model

  Build a quantitative customer funnel — visitors → signups → activations → paid → retained — with per-step conversion rates and end-to-end yield. Hands off to mvp-planning/funnel-instrumentation-spec to translate stages into events.

• get-keep-grow-design

  Walk the three stages of customer relationships — Get (acquisition), Keep (retention), Grow (expansion). Each stage has its own metrics and investments. Writes 06-relationships-channels/get-keep-grow.md.

• hypothesis-falsifiability-check

  Audit a hypothesis for the four failure modes — no falsifier, no measurement, no threshold, no timeframe. Blocking. --force overrides and is logged.

• hypothesis-register

  Maintain the canonical list of venture hypotheses — id, BMC cell, statement, status (open/accepted/refuted/superseded), falsifier, measurement, threshold, timeframe, evidence. Add, update, flip status. Single source of truth for what we believe.

• interview-analyse

  Aggregate findings across all interviews in a segment — count confirms/refutes per hypothesis, surface emergent themes, flag outliers. Outputs interview-summary.md plus proposed hypothesis-register updates for the user to approve.

• interview-guide-build

  Generate a customer-discovery interview guide tied to the hypotheses being tested. Constrained to ≤ 30 minutes, open questions, the standard "why / why not / who else / can we follow up" closers. Writes segments/<slug>/interview-guide.md.

• interview-log

  Capture an interview as raw notes plus structured extraction — which jobs/pains/gains were confirmed, new, or refuted, and which hypotheses moved. Append-only — once filed, an interview is never edited.

• learning-card-build

  Generate a 4-part Strategyzer learning card — we tested / observed / learned / will now — for a concluded test card. Proposes the matching hypothesis-register flip.

• migration-plan

  Sequence migrations for a green-field MVP — bootstrap, seed data, RLS, edge functions, types regeneration. Pairs with supabase-schema-design. Optionally orchestrates apply_migration calls via the connector-confirmation flow.

• mvp-analytics-plan

  Translate mvp-metrics into an analytics implementation — event taxonomy, properties, identity model, tool selection (PostHog / GA4 / Plausible / Mixpanel). Writes 09-mvp/analytics/events-spec.md.

• mvp-build-plan

  Translate mvp-spec.md into a sprint plan — features → tickets → estimates, paired to hypotheses. Delegates to software-development/plan-orchestrator (upstream) for the parallel multi-agent coverage check when an audit-grade plan is needed.

• mvp-deploy-plan

  Plan deployment for the selected MVP stack, including Vercel or Cloudflare where applicable. Do not provision or deploy.

• mvp-feasibility

  Surface technical, regulatory, and resource feasibility risks. Reads outputs of mvp-tech-plan, mvp-schema-plan, mvp-deploy-plan plus optional Supabase/Cloudflare MCP quota probes. Writes 09-mvp/feasibility.md.

• mvp-metrics

  Define the success metrics per hypothesis the MVP tests. No hypothesis without a metric, no metric without a threshold, no threshold without a timeframe. Writes 09-mvp/mvp-metrics.md.

• mvp-open-questions

  Aggregate unresolved MVP planning items with source references. Refresh the report explicitly without requiring external hooks.

• mvp-schema-plan

  Translate MVP scope and VPC into a database schema. Sequenced delegation to data-model-from-vpc → supabase-schema-design → optionally upstream database-design/postgres-schema-audit for QA. Thin orchestrator.

• mvp-scope

  Define the smallest feature set that lets us test the *primary* hypothesis. Forces a cut/keep/maybe classification of every candidate feature. Blocking — refuses without a green customer-discovery-status (override with --force, logged).

• mvp-tech-plan

  Translate the MVP scope into a tech plan — sequenced delegation to tech-stack-recommender → architecture-design → adr-writer. Thin orchestrator. Outputs reside under 09-mvp/.

• mvp-type-select

  Choose between pre-order, audience-building, show-and-tell, or partial-product MVP types. Each is a real MVP; they test different things. Updates 09-mvp/mvp-spec.md.

• paper-prototype

  Generate a paper-prototype script — the screens / states / interactions sketched as text or pen-and-paper, and a guided run-through script. Cheaper to be wrong on paper. Writes 08-prototype/paper/<slug>.md.

• phase-router

  Recommend the next 1–3 actions for the venture based on its current state. Reads the index, hypothesis register, learning cards, and open questions. Outputs a prioritised plan with the skill name to invoke for each action. Read-only.

• pitch-1min-build

  Generate a 1-minute pitch from the current BMC + UVP + MVP scope. Outputs a script (~150 words) plus a slide-skeleton outline. Suitable for pitch competitions, investor coffee chats, customer cold-opens.

• pivot-refine-log

  Append a pivot or refine entry to 07-validation/pivot-refine-log.md. Forces the canonical entry shape — trigger evidence, decision, what changed, what was kept, new version pointers. Surfaces stale pages that need follow-up.

• product-channel-fit-check

  Ensure the product type and chosen channel are coherent — SaaS sold via retail is incoherent; mass-market consumer good sold by inside sales is incoherent. Outputs a fit verdict per primary channel.

• prototype-feedback-collect

  Standardised template for capturing prototype-test feedback sessions. Writes one file per session under 08-prototype/feedback/<prototype-slug>-<NNN>.md. Append-only — sessions are immutable once filed.

• prototype-vs-mvp-distinguish

  Classify an artifact against the 5-dimension prototype-vs-MVP rubric (audience, fidelity, scope, environment, what-it-proves). Refuses to label a prototype as an MVP. Blocking — `--force` overrides, override is logged.

• six-ways-to-innovate

  Apply Strategyzer's six lenses (customer needs, value, product, experience, relationship, channel) to the current customer profile and value map. Generates candidate VP refinements as proposed hypotheses, not facts.

• supabase-schema-design

  Design the full Supabase schema for the MVP — tables, columns, FKs, indexes, RLS policies, RPC functions, triggers. Uses the Supabase MCP for live introspection when available; gracefully degrades. Mutations gated by connector-confirmation idiom.

• swot-build

  Generate a SWOT for a top competitor. Enforces the internal (S/W) vs external (O/T) distinction. Writes 04-competitors/swot/<competitor-slug>/README.md.

• tech-stack-recommender

  Compare viable technology stacks for a startup MVP using the hypothesis, existing system, team, budget and data constraints; recommend the simplest justified option.

• test-card-build

  Generate a 5-part Strategyzer test card — we believe / to verify we will / measure / are right if / threshold. Writes 02-customer-discovery/test-cards/TC-NNN.md.

• uvp-statement

  Produce the standard UVP statement — "For [customer] who [need], our [product] is a [category] that [unique benefit]. Unlike [competitor], we [differentiator].

• value-map-build

  Build the left half of the VPC for one segment — products & services, pain relievers, gain creators. Maps each reliever to a prioritised pain and each creator to a prioritised gain. Writes 03-value-proposition/vpc-<segment>-vN.md.

• venture-handoff-doc

  Generate a one-page brief summarising the venture — problem, segment, UVP, validated hypotheses, MVP scope, open risks, current phase. Suitable for a new team member, investor, or future-you. Read-only; writes a single handoff.md.

• venture-init

  Create a self-contained venture workspace, or reuse a verified existing one, without overwriting files or requiring Memex.

• venture-status

  Print a human-readable snapshot of the current venture — phase, hypothesis tallies (open/accepted/refuted), interviews logged per segment, BMC version, outstanding test cards, recent log activity. Read-only — never mutates state.

• vision-sketch

  Capture the venture's initial vision as a sketch (not a spec). Forces answers to three questions — top customer problems, how the idea helps, day-in-the-life before vs after — and writes them to 00-vision/vision-sketch.md and day-in-life.md.

• vpc-fit-check

  Cross-check that every prioritised pain on a segment profile has at least one pain reliever and every prioritised gain has at least one gain creator. Appends a fit-report block to the latest VPC. Sets status to active if fit; draft if partial.

• vpc-version

  Snapshot the current VPC and bump to v(N+1) as customer feedback refines it. Tracks the diff so the evolution story is auditable. Sets the prior version's status to superseded.

Subagents (9)

• business-model-orchestrator

  Use this agent to drive the canonical BMC sequence — bmc-build (v1), bmc-link-vpc to wire VPCs into the BMC, bmc-revenue-cost-sketch for envelope economics, bmc-update on hypothesis flips, bmc-front-back-split for sequencing. Read-only; produces a step plan.

• competitor-analysis-orchestrator

  Use this agent to drive the canonical competitor analysis sequence — competitor-discover → competitor-table-build → top-3 swot-build → competitor-bmc-shadow for top-2 → uvp-statement → competitor-insights. Read-only; produces a step plan grounded in current competitor coverage.

• customer-discovery-orchestrator

  Use this agent to drive the canonical customer-discovery sequence for a segment — segment-define → profile-build → early-adopter-profile → interview-guide-build → interview-log loop → interview-analyse → check the four-question gate. Read-only on the venture; produces a prioritised plan grounded in

• experimentation-orchestrator

  Use this agent to drive the canonical Ch. 5 experimentation loop — prioritise hypotheses, design experiments, build test cards, run-track, build learning cards on conclusion, propose hypothesis flips. Read-only; produces a step plan.

• mvp-planning-orchestrator

  Use this agent to drive the MVP planning sequence — mvp-scope → mvp-type-select → mvp-metrics → mvp-tech-plan → mvp-schema-plan → mvp-deploy-plan → mvp-analytics-plan → mvp-feasibility → mvp-build-plan → pitch-1min-build. Read-only on venture state; produces a step plan grounded in the customer-disc

• prototyping-orchestrator

  Use this agent to drive the prototype-workflow sequence — divergent-ideate → converge-ideas → paper-prototype → prototype-feedback-collect (×N) → optionally digital-prototype. Read-only; produces a step plan grounded in the venture's hypotheses and segment state.

• relationships-channels-orchestrator

  Use this agent to drive the Ch. 8 sequence — get-keep-grow design, channel-select, product-channel-fit-check, funnel-model, churn-model. Read-only; produces a step plan grounded in the venture's BMC and segment definitions.

• value-proposition-orchestrator

  Use this agent to drive the canonical VPC sequence for a segment — value-map-build → vpc-fit-check → optional six-ways-to-innovate → vpc-version when feedback warrants. Read-only; produces a step plan grounded in the segment's customer profile and existing VPCs.

• venture-core-orchestrator

  Use this agent to drive the canonical "land a new venture" sequence — confirm the workspace is initialised, write the vision sketch, seed the hypothesis register, scaffold the first segment, generate an interview guide. Read-only on the venture state during planning; the actual writes happen via the

Similar plugins