Parent: #318
Blocked by: owner authorization, #321, and #322
Goal
Validate the CI-reporting wedge with reversible, maintainer-respectful external pilots. The objective is learning and real consumer evidence, not indiscriminate adoption PRs.
Target selection
Identify five active, permissively licensed Python repositories that:
- use GitHub Actions and emit JUnit XML, coverage, Ruff, ty, SARIF, or comparable structured artifacts;
- have a clear reporting gap or an explicit custom-report need;
- have recent maintainer activity and contribution guidance that permits the proposed change;
- can accept a small, reversible workflow-only PR; and
- do not require Kida to replace an embedded Jinja/IaC ecosystem.
Pilot protocol
For each selected target, record before outreach:
- repository, immutable base revision, maintainer/contribution constraints, current workflow, input artifact, and why the Kida Action is relevant;
- a minimal proposed workflow diff, exact Action version, permissions, and rollback;
- the corresponding evidence card/fixture; and
- an explicit downstream-pilot classification under
docs/downstream-pilot-policy.md.
After owner authorization, submit at most one narrow PR per target. Do not modify unrelated workflow, test, permissions, or reporting behavior.
Acceptance and proof
- Five candidates are documented before any PR is opened.
- Each submitted PR links to its exact fixture and discloses runner/platform and permission assumptions.
- Track response, requested changes, merge/rejection, observed runs, and reasons without treating rejection as product failure.
- For any accepted pilot, record 14 days of real runs or the policy’s exact applicable observation alternative, plus a maintainer-feedback summary.
- Feed discovered gaps into separate bounded issues; do not silently widen Kida’s scope in the pilot PR.
Boundaries
- No third-party PR, maintainer contact, issue comment, or external release without fresh owner approval for that campaign.
- No use of private data, credentials, or write tokens beyond the target repository’s documented contribution path.
- No claim of endorsement, adoption, or performance improvement until the target maintainer and evidence support it.
- Do not turn report-only downstream observation into a required gate.
Stewards and collateral
- Stewards: docs, github, action, templates, markdown, schemas, utils, public, tests, examples/site; other consulted stewards only if a target exposes their surface.
- Collateral: per-target pilot record, PR link, source revision, fixture link, observed-run evidence, and a public case study only with maintainer-appropriate attribution.
- Downstream-pilot requirement: required for accepted downstream-observable workflow use; use the policy’s consumer record exactly.
Blocking decision
This task is intentionally blocked. Creating this issue does not authorize contacting or changing any third-party repository.
Parent: #318
Blocked by: owner authorization, #321, and #322
Goal
Validate the CI-reporting wedge with reversible, maintainer-respectful external pilots. The objective is learning and real consumer evidence, not indiscriminate adoption PRs.
Target selection
Identify five active, permissively licensed Python repositories that:
Pilot protocol
For each selected target, record before outreach:
docs/downstream-pilot-policy.md.After owner authorization, submit at most one narrow PR per target. Do not modify unrelated workflow, test, permissions, or reporting behavior.
Acceptance and proof
Boundaries
Stewards and collateral
Blocking decision
This task is intentionally blocked. Creating this issue does not authorize contacting or changing any third-party repository.