chore(project): audit readiness and consolidate browser fixtures

[TusanHomichi]

Primary Issue

Closes #53
Also closes #57
Refs #44

Problem And Outcome

Fresh contributors received contradictory status and repeated guidance; CI
never exercised the shipped browser workflows. Five specs also duplicated
server management, accepted any healthy process on a fixed port, and did not
await shutdown. This change makes context task-directed and gives test
installations one lifecycle owner.

Changes

• Reduce AGENTS.md from 822 to 605 words while retaining the project contract.
  CONTRIBUTING.md owns workflow detail; docs/development.md maps tasks to
  source, ADRs, formats, and tests. Tool entrypoints remain thin.
• Correct current-versus-planned claims and the existing review finding:
  snapshot-bound authorization is a packet-path property, not a universal
  guarantee. Clarify ADR 0011's already-enforced ASCII member-name restriction.
• Replace duplicated browser startup/setup-code/teardown with a shared
  Playwright fixture. Each test gets an OS-assigned loopback port and scratch
  installation; readiness requires the child to announce its bound listener.
  Shutdown is awaited before scratch data is deleted on both pass and failure.
  Startup diagnostics exclude setup codes.
• Add regression checks for missing binaries, early exit, occupied listeners,
  distinct installation ports, shutdown, cleanup, and diagnostic redaction.
• Include browser/config TypeScript in npm run check with explicit Node type
  declarations. Add the explicit binary build and browser suite to CI.
• Add the preview runbook and dated audit. The reproduced doctor mutation
  remains a separate repair in storage(doctor): diagnose without changing persisted journal mode #56; the guides now qualify that contract.

Scope

• In scope: contributor/agent guidance, continuation audit, test-fixture
  ownership, typechecking, existing-workflow CI coverage, preview runbook.
• Out of scope: application behavior or schema changes, milestone
  implementation, remaining linked defects, or preview deployment changes.

Verification

• git diff --check
• Local Markdown target check: 54 relative links across 34 Markdown files,
  with no missing file targets.
• web/: npm ci
• web/: npm run check (Svelte and browser TypeScript pass)
• web/: npm run build
• cargo fmt --check
• cargo clippy --workspace --all-targets -- -D warnings
• cargo test --workspace (118 integration tests passed)
• cargo build -p consolebook-server
• web/: CONSOLEBOOK_E2E_CHROMIUM=/usr/bin/google-chrome npm run e2e
  (five original scenarios and three new fixture tests: 8 passed in 29.4s)

The first explicit browser typecheck failed with missing Node declarations
(TS2307); the new development dependency and check configuration correct it.
The preexisting low-severity npm advisory chain remains tracked in #54.
The doctor reproduction used a scratch toy database; its expected
missing-schema exit is distinct from the observed journal-mode mutation.

CI at 3a6cd8e passed the added browser stage before the fixture consolidation.
CI at 3d87379 also passed the consolidated fixture and browser TypeScript checks.
Final head 424f01e only links the remaining audit ownership work to #58 and #59;
its full gate is https://github.com/FieldmouseWorks/consolebook/actions/runs/33928869139.
All fixtures are invented. No production recovery claim.

Review And Merge Notes

Review the authority/context split, snapshot-authorization qualification,
fixture resource ownership/failure paths, and browser CI step. The dated
audit lists unresolved work without claiming those contracts repaired.
The public preview remains on its separately installed release binary.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-04T23:09:03.204172Z	3d87379	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f369981d4b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".