# Partner Onramp

> Check the partner-developer journey on a B2B SaaS platform, from signup to the first submitted app, and find where
> partners stall. Describe the program, list the journey's stages with their gates, owners and metrics, the
> post-launch thresholds and this quarter's funnel counts; the browser ranks the entry gate, provisioning, education
> and the stacked support ladder, audits the journey for the skill's failure modes and reads the four funnel
> checkpoints against its thresholds. Metered runs write the onboarding design, and the fixes. Built on the SkillSafe
> skill @samber/partner-app-onboarding, from samber/developer-platform-skills (MIT), by Samuel Berthe.

https://partner-onramp.skillsafe.ai/

## What is free
- The rungs: self-service signup, an agreement-split seller track, or an application review (weeks to a first app, sensitive data); a dedicated dev tenancy at signup, the partner's own account (safe actions only), or production-mirror sandboxes; an optional, credentialed or mandatory education posture; a stacked support ladder of forum, chat, office hours and named managers for commercial ISVs.
- The journey audit: credentials issued after a review gate, pricing asked before submission, certification gating submission, post-launch thresholds missing from the onboarding docs, stages with no owner or metric, an unpublished review checklist.
- The funnel: signup to provisioned, provisioned to first API call, first call to first submission, submission to approval - the biggest leak, over ten minutes to a first call, a first-review fail rate of 30% or more with no validator, and time to first submission far beyond review time.
- The gates: building decoupled from approval, commercial decisions at launch, thresholds disclosed, stages owned and measured, the funnel counted; plus a journey table, a funnel CSV, the checks (.csv) and everything (.json).
- The loop: save an onramp each quarter; the funnel table then shows each checkpoint's change in points against the latest earlier save for the same platform.

## The metered lanes
Two tasks on gpt-terra, metered per run, routed by `task`. Input string fields: `task` (design or fixes), `program`, `journey`, `carry` (fixes only: the design's decisions) and `facts`.
- `design` replies with `summary`, `stages`, `rungs`, `promotions`, `commercial`, `funnel`, `disclosure` and `caveats`.
- `fixes` replies with `summary`, `fixes` (each issue with its fix, owner and metric, biggest leak first) and `caveats`.
The page holds every reply to its own facts: the ranked rungs, named and owned stages, credentials before app review, pricing at launch, four checkpoints and a self-defined headline metric; every issue in the page's order with the specific change it needs.

## Links
- API tutorial: https://partner-onramp.skillsafe.ai/api.html
- Source skill: https://skillsafe.ai/skill/@samber/partner-app-onboarding