developer-champion-programs

---
name: developer-champion-programs
description: 'Use when designing a consent-aware advocate or champion program with qualification, benefits, and opt-in records. Not for unsolicited contact, quotas, or forced referrals.'
disable-model-invocation: true
---

# Developer champion programs

Design an advocate, champion, insider, or ambassador program that people want to join, and that never pressures anyone. The result is a program design in chat. No one is contacted or enrolled.

## Consent is the whole foundation

Every behavioral claim in the design traces to evidence, and every public use of a person (name, quote, or likeness) has a recorded permission behind it. No testimonial is ever invented. The following are refused outright, no matter how the request is phrased: unsolicited contact, mandatory advocacy, content quotas, commission or MLM framing, forced referrals, automatic swag, and any identity use without consent.

Public advocacy appearances themselves belong to `developer-advocacy`; this skill owns the program structure and its consent contract, which take different inputs and produce a different output.

## What you need before starting

- `goal` (required): what the program exists to do.
- `evidence` (optional): behavioral signals already observed (contributors, helpers, writers), with sources. No signal is ever invented.
- `constraints` (optional): budget, staffing, and brand limits.

## Shaping the program

Work through these in order:

1. Define qualification signals from observed behavior only, with sources. Anything not yet observed is labeled proposed-not-observed.
2. Set benefits proportional to contribution and capacity (access, recognition, learning, influence on roadmap input). Each benefit names its cost and its owner.
3. Set expectations without quotas: what champions may do, never must do. Mandatory advocacy and content quotas are refused explicitly, and the final design contains neither.
4. Write recognition rules: public use of a name, quote, or likeness requires recorded permission, and opt-out is immediate and complete.
5. Draw referral boundaries: invited, never extracted. No misleading commission framing, no forced referrals. Invitation-only and honest.
6. Open the opt-in ledger: who may enroll, how consent is recorded, and how withdrawal works. All three must be specified.

## When the ground is shaky

- No observed behavioral evidence: design the pilot around proposed signals flagged as such. Do not invent current champions.
- Budget or staffing too small for the ambition: cut benefits to fit and name the cut.

## What you hand back

A program design: goal, qualification signals with evidence, benefits with costs, expectations without quotas, recognition and permission rules, contributor routes, referral boundaries, and the opt-in ledger specification. Every behavioral claim carries evidence; every public use carries a permission record.

• SKILL.md 3 KB
• agents/openai.yaml 1 KB