docs: align contribution and bot review with change impact

[tt-a1i]

Problem and value

Reviewers currently rediscover impact and acceptance boundaries, while repository-only prose and focused test corrections inherit a local full-suite requirement. Bot checks can also duplicate requests: on PR #351, scope assessment requested archive evidence while Validation Evidence had already inspected that archive.

Bound review by changed behavior and callers, settle scope before expensive integration work, and re-review new changes plus unresolved findings. CONTRIBUTING.md owns author requirements; REVIEWING.md owns reviewer procedure; the shorter template collects decision-relevant evidence. Align CodeRabbit with these same rules. Refs #344.

Stability impact

• Repository review policy only: CONTRIBUTING.md, REVIEWING.md, the PR template, and .coderabbit.yaml. No runtime, packaged Skill, schema, CI workflow, or generated-source changes relative to main.
• Keep full local checks for runtime, schema, packaged authoring behavior, generated content, and shared test infrastructure. Repository-only prose/policy and focused test-only changes may explain targeted evidence. Remote CI and branch protection remain unchanged.
• Preserve compatibility, authored geometry, stable diagnostics, and separate browser/perceptual evidence. Historical evidence retains its revision.
• Bot: keep chill, ready-PR/incremental review and two warning checks; include REVIEWING.md; disable overlapping built-in issue assessment and in-progress fortune messages. Check the PR diff rather than checkout dirtiness, accept evidence already supplied, and distinguish author work from maintainer CI approval. Narrow fixes can use reproduction; new contracts/defaults need agreed scope.
• Document incremental/full review, pre-merge rechecks after evidence-only edits, and pause/resume. Bot guidance is probabilistic and advisory; it cannot grant merge authority.
• Rollback: revert these four policy/config changes. No data migration or artifact rebuild is needed.

Tests run

Comparison base: 06bd6fea5752bc06b8170a1f09085c798df5aa13. Candidate: a3bdf15 (full SHA in PR head).

On the candidate content with Node 22.23.2:

• node --test archify/test/community-proof-intake.test.mjs: 3 passed, 0 failed, 0 skipped.
• git diff --check origin/main: clean.
• PyYAML 6.0.3 + jsonschema 4.26.0 validation against the official v2 schema: zero errors; all four guideline paths exist. Ten relative Markdown links/anchors resolve.

Reproduce schema validation (Python environment with PyYAML/jsonschema):

curl -fsSL https://coderabbit.ai/integrations/schema.v2.json -o /tmp/coderabbit-schema.json
python - <<'PY'
import json, pathlib, yaml
from jsonschema import Draft7Validator
config = yaml.safe_load(pathlib.Path('.coderabbit.yaml').read_text())
schema = json.loads(pathlib.Path('/tmp/coderabbit-schema.json').read_text())
errors = list(Draft7Validator(schema).iter_errors(config))
for error in errors:
    print(error.json_path, error.message)
assert not errors
for path in config['knowledge_base']['code_guidelines']['filePatterns']:
    assert pathlib.Path(path).is_file(), path
print('Schema errors: 0; guideline paths: 4')
PY

No local renderer suite: the four changed files do not affect renderer, package or site inputs. Current-head remote checks and bot review must be read separately; older CI is not a current pass.

Historical procedure pilots at the previous policy candidate e9a6fe67545a0433c72c49facdfaf851ed339292 (reused design evidence, not new-head executions):

• PR feat: support per-lane workflow heights #255 at ab4908b1618002f3c1372303d1106f0c56f4b4ca: contract/shared change needs a scope decision on automatic v2 lane sizing versus public lane.height semantics before another full implementation review. No new browser/full-suite claim.
• PR test: load workflow migration preload via a file URL #319 at 16ef0cd3e6c4031be906f89de1d92e0e46c069b7: focused test-only change fits targeted verification; Windows base-fail/head-pass remained author-reported and remote CI required action. No merge acceptance inferred.

Current policy review scenarios: repository prose needs content/link checks; shared geometry needs caller tracing and comparable relevant artifacts; fork CI awaiting approval needs maintainer action; reused archive evidence should not be requested again solely because a binary diff is excluded. These are manual policy-consistency checks, not proof of future LLM behavior or measured review-time savings. Observe the first 5–10 reviewed PRs before expanding enforcement.

Visual evidence

Not applicable: repository review policy only; no diagram, Viewer or public-page appearance/interaction changes.

Generated artifacts

None. These root/GitHub policy files are outside packaged archify/ and generated site inputs; ZIP and Gallery sources are unchanged.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Advanced

Run ID: 738403bf-74af-4b0d-af0d-929408d77799

📥 Commits

Reviewing files that changed from the base of the PR and between 06bd6fe and a3bdf15.

📒 Files selected for processing (4)

• .coderabbit.yaml
• .github/PULL_REQUEST_TEMPLATE.md
• CONTRIBUTING.md
• REVIEWING.md

---

📝 Summary

Summary

Updates repository review policy and CodeRabbit configuration to align review scope and evidence with change impact.

• Defines incremental and full review procedures in REVIEWING.md and CONTRIBUTING.md.
• Adds guidance for scope assessment, evidence reuse, re-review boundaries, pause/resume handling, and maintainer CI approval.
• Updates the PR template to require problem, stability, test, visual evidence, and generated-artifact details.
• Configures CodeRabbit to review the PR diff, consolidate duplicate requests, and avoid overlapping issue assessment and progress messages.

Compatibility impact

Changes are limited to repository documentation, the PR template, and .coderabbit.yaml. Runtime, packaged Skill, schema, CI workflow, generated-source, renderer, browser, and public-page behavior are unchanged.

Validation

The author reports focused tests, whitespace checks, CodeRabbit schema validation, guideline-path checks, and Markdown link resolution. These checks were not independently observed here. No browser or perceptual acceptance is claimed.

Walkthrough

Changes

Review process and contribution guidance

Layer / File(s)	Summary
Review policy and reviewer procedure .coderabbit.yaml, REVIEWING.md	Review instructions now use revision-aware scope, evidence classification, ownership, and request consolidation. REVIEWING.md defines the review and re-review procedure.
Contributor workflow and pull request contract .github/PULL_REQUEST_TEMPLATE.md, CONTRIBUTING.md	Contribution guidance and the pull request template now capture scope, impact, compatibility, tests, visual evidence, artifacts, and integration status.
Impact evidence and artifact handling CONTRIBUTING.md	Guidance now defines evidence by impact, runtime and browser checks, package reproducibility, generated outputs, and published artifact immutability.
Automated review operations CONTRIBUTING.md	The automated review section documents warning checks, review commands, fork-CI approval, and pause/resume operations.

Estimated code review effort: 3 (Moderate) | ~20 minutes

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}