Skip to content

Latest commit

 

History

History
114 lines (77 loc) · 9.05 KB

File metadata and controls

114 lines (77 loc) · 9.05 KB

Changelog

All notable changes to launchpad-toolkit are documented in this file.

The format is based on Keep a Changelog. During the v0.x prototype phase the plugin does not follow strict semver — breaking changes may occur in minor releases as methodology iterates. Strict semver compliance begins in v1.0.0.

No unreleased changes. v0.5 shipped the final v0.x scope item (dojoos-api-consumer agent). Next milestone is v1.0 stability pass.

0.5.0 — 2026-04-15

Added

  • Agent dojoos-api-consumer — spec-driven enrichment layer that consumes the DojoOS OpenAPI spec (shipped via DOJ-3170 at https://docs.dojocoding.io/openapi.yaml / local mirror in dojo-documentation/public/openapi.yaml). Reads the spec once per session, caches it, resolves skill-facing semantic aliases (e.g. list_candidate_pool, get_investor_database, sync_stage) to OpenAPI operationIds, and executes curl-based HTTP calls with Bearer JWT auth. Returns a structured YAML contract with status codes:
    • LIVE_DATA — endpoint reachable + returned data
    • NOT_IMPLEMENTED — endpoint in spec but 404/501/timeout
    • SPEC_GAP — operation not in spec at all (seeds a feature-to-spike proposal)
    • AUTH_REQUIRED — token missing/expired
    • RATE_LIMITED — 429 with Retry-After
    • ERROR — network/parse/unknown 5xx
  • Agent never fabricates API responses — graceful degradation is the default behavior, so every calling skill continues to work standalone when the requested endpoint is missing.
  • Agent defaults to staging (pphagffyuibcfulgrpjb.supabase.co/functions/v1) — never hits production without explicit opt-in via DOJOOS_API_BASE_URL.
  • Security posture: bearer token never logged, never persisted, redacted from all output; non-DojoOS redirects treated as errors.

Changed

  • plugin.json 0.4.0 → 0.5.0
  • Skills cofounder-matching, investor-matching, stage-tracker — "Integración con DojoOS (future)" sections updated to reference the now-available dojoos-api-consumer agent (1-2 sentences each, no methodology changes).
  • README.md — adds v0.5 section, lists the agent alongside the 8 skills + 1 command, clarifies that all 4 v0.x releases have shipped.
  • Keyword dojoos-api-consumer added to plugin.json.

Rationale

v0.5 was originally scoped as "blocked on @garbanzo's API implementation". Re-evaluation revealed two flawed assumptions:

  1. The OpenAPI spec already exists (DOJ-3170 closed 2026-04-14) — the agent can be implemented against contract, not infrastructure.
  2. The agent does not require every endpoint to be live — it handles SPEC_GAP (no endpoint) and NOT_IMPLEMENTED (endpoint stubbed) as first-class outcomes, so it ships useful on day one and becomes progressively more useful as @william + @garbanzo expose Launchpad-specific endpoints.

Today the agent unlocks LIVE_DATA for the 12 general-purpose endpoints already in the spec (Agent chat, course progress, certificates, B2B leads, billing, role switch, admin analytics, media tokens, email). Launchpad-specific operations (cofounder pool, investor DB, demo-day queue, stage sync, Dojo Score) return SPEC_GAP today — by design — and each SPEC_GAP auto-seeds a SPIKE proposal for William via the feature-to-spike loop.

0.4.0 — 2026-04-16

Added

  • Command /launchpad-toolkit:propose-spike — thin wrapper sobre el skill feature-to-spike con defaults pre-configurados para William Ugalde (DojoOS Launchpad pillar owner):
    • Auto-assignee: William (Linear ID 8f14370d-3602-49e3-81f2-eeb05b965687)
    • Auto-parent: DOJ-3189 (launchpad-toolkit tracking SPIKE)
    • Auto-labels: Spike, Explore, methodology-prototype
    • Auto-team: DojoOS; Auto-project: Launchpad; Auto-state: Triage
    • Accepts optional $ARGUMENTS como seed del "concrete pattern" en FTS-1
    • Direct file via Linear MCP (mcp__linear-server__save_issue) o fallback markdown copy-paste
  • Documentation of when NOT to use the command (feature requests, bugs, plugin internals — estos van por otras vías)

Changed

  • plugin.json 0.3.0 → 0.4.0
  • CHANGELOG v0.5 roadmap simplificado: solo dojoos-api-consumer agent. Descubrimiento: la OpenAPI spec YA existe en dojo-documentation/public/openapi.yaml (shipped via DOJ-3170). El agente puede implementarse contra el spec SIN depender de @garbanzo para nueva infraestructura — lo que bloquea es solo endpoint availability por-endpoint.

0.3.0 — 2026-04-16

Added

  • Skill cofounder-matching — rubric-based co-founder matching methodology con 6-axis framework (domain fit, skill complementarity, values alignment, equity expectations, time commitment, track record). Weighted scoring ajustado por stage. Genera scorecard + gap analysis + interview questions + ranked shortlist. Kill-switches override score. Standalone; sync con DojoOS API cuando el agente dojoos-api-consumer esté disponible.
  • Skill investor-matching — investor fit scoring con 5-axis framework (stage fit, check size, thesis alignment, geography/vertical, value-add depth). Weighted scoring configurable por priority (fundraising-first / value-add-first / stealth-strategic). Genera fit-scorecard + intelligence notes + customized outreach + target list ranked + pipeline tracker.
  • Skill demo-day-prep — preparation workflow para Demo Day events (YC, Techstars, etc.). 4 artefactos: application write-up, 10-slide deck outline con speaker notes, 3-min pitch script con timing markers, 50+ Q&A bank categorizado. Incluye rehearsal best practices (5-5-5 rule) y anti-patterns. Standalone; schema-compat con DojoOS Demo Day queue futuro.
  • Skill stage-tracker — milestone tracking across the 6 DojoOS Launchpad stages (Ideation → Formation → MVP → Traction → Funded → Scaling). Exit criteria per stage + Dojo-Score-compatible 5-axis scoring (Team, Product, Traction, Market, Business Model; 0-100 total). Genera current-stage + milestones + blockers + dojo-score + history. Evidence-over-claim principle.

Changed

  • plugin.json version bump 0.2.0 → 0.3.0
  • README updated to reflect 8 skills total (up from 4)
  • Roadmap simplified: v0.4 = dojoos-api-consumer agent + /propose-spike command (v0.3 is the last "skills-only" release)

Rationale

Previously v0.3 skills were blocked on DojoOS API via @garbanzo. Re-evaluation revealed that skills themselves are methodology (rubrics, scoring, workflows) — they do NOT require live API data. Only the dojoos-api-consumer agent (which fetches live data to enrich skills) is API-blocked. Skills ship as standalone; agent is v0.4 pending API.

0.2.0 — 2026-04-16

0.2.0 — 2026-04-16

Added

  • Skill cap-table-builder — constructor de cap table inicial para single venture. Soporta founders + option pool + advisor pool + SAFEs + convertible notes. Genera cap-table.md + vesting-schedule.md + safe-conversion-modeling.md con modeling de 3 scenarios (conservative/expected/optimistic) de next priced round. Cross-reference a venture-studio-toolkit:sweat-equity-agreement para 83(b) tratamiento diferenciado por entity type (C-Corp restricted stock / LLC profits interest / LLC capital interest).
  • Skill founder-documents — generador de stack legal founder. 6 documentos: Founder Stock Purchase Agreement (o Operating Agreement LLC), IP Assignment, Vesting Schedule Exhibit, Advisor Agreement (FAST framework), SAFE (post-money YC standard), Term Sheet (NVCA Series Seed). Cada output incluye [TO BE FILLED BY LAWYER] markers y lawyer review checklist. STRONG legal disclaimer en todos los outputs.

Changed

  • plugin.json version bump 0.1.0 → 0.2.0
  • README references updated to reflect 4 skills total (up from 2)

0.1.0 — 2026-04-16

Added

  • Initial scaffold (DOJ-3221, parent SPIKE DOJ-3189)
  • plugin.json + README.md + BSL-1.1 LICENSE
  • Skill startup-intake — AI intake interview producing startup-profile.md artifact, schema-compatible with DojoOS Launchpad Startup Profile
  • Skill feature-to-spikedifferentiator — transforms dog-food methodology learnings into Linear SPIKE issues for William Ugalde (Launchpad pillar owner)
  • Reference references/productization-workflow.md — 5-step loop documenting methodology → SPIKE → DojoOS feature flow

Known limitations (v0.1)

  • Only 2 of 8 planned skills implemented; remaining 6 deferred to v0.2/v0.3
  • dojoos-api-consumer agent not implemented (blocked on DojoOS API by Daniel Garbanzo)
  • Bilingual output framework documented but implementation per-skill is roadmap