G
Gait

Ecosystem Contribution Funnel

Ecosystem Contribution Funnel

Use this path to propose a community adapter or skill and get it listed.

1) Open The Correct Proposal

  • adapter proposal: .github/ISSUE_TEMPLATE/adapter.yml
  • skill proposal: .github/ISSUE_TEMPLATE/skill.yml

Required in proposal:

  • problem solved and target runtime/framework
  • minimal runnable example path
  • fail-closed behavior (allow vs non-allow execution)
  • deterministic output paths + test plan

2) Respect v2.3 Lane Governance

Blessed lane in v2.3:

  • coding-agent wrapper + GitHub Actions regress template

No new official lane is merged without scorecard evidence:

python3 scripts/check_integration_lane_scorecard.py \
  --input gait-out/adoption_metrics.json \
  --out gait-out/integration_lane_scorecard.json

Gate for official-lane expansion:

  • selected score >= 0.75
  • confidence delta >= 0.03

If threshold is not met, lane can still ship as community/experimental.

3) Add Or Update Community Index Entry

Edit docs/ecosystem/community_index.json:

  • add one entry with unique id
  • point repo to public GitHub URL
  • set source to community unless maintained in this repo
  • set initial status to experimental for new submissions

Validate locally:

python3 scripts/validate_community_index.py

4) Prove Contract Compatibility

Adapter submissions should include:

  • intent evaluation via gait gate eval (or MCP boundary equivalent)
  • deterministic trace/run artifact paths
  • explicit non-allow non-execution behavior

Validation commands:

make lint
make test-adoption
make test-adapter-parity

Skill submissions should include:

  • provenance metadata (source, publisher, digest/signature where applicable)
  • wrapper-only orchestration (no policy parsing/evaluation logic outside Go)

Validation commands:

make lint
make test-skill-supply-chain

5) Review And Listing

A contribution is listed in docs/ecosystem/awesome.md once:

  • index entry validates in CI
  • reviewer confirms deterministic/no-bypass behavior
  • install + quickstart commands are documented

6) Release Automation Artifact

Before tagged release, generate deterministic ecosystem release notes:

python3 scripts/render_ecosystem_release_notes.py

Output:

  • gait-out/ecosystem_release_notes.md