Tasks: write the-loop's artifacts for a human reader

Derived from design.md and testing-plan.md. TDD invariant (tdd.mode: standard): the parity test lands first and red.

Task list

• [x] 1. Land cli/tests/test_writing_parity.py red
  • P1–P4 (originally P1–P6, before budgets were removed — task 8)
  • Depends on: none · Requirements: R5.2, R5.3
  • Test: T1 — pytest cli/tests/test_writing_parity.py (red: no skill yet)
• [x] 2. Author skills/writing/SKILL.md and reference/tells.md
  • Spine, density test, diagram-first, carve-out, protected content, revise pass
  • Depends on: 1 · Requirements: R1.1, R1.2, R1.3, R3.1, R4.1, R4.2
  • Test: T1 — P1, P4 (green)
• [x] 3. Add userInteraction.writingStyle to the schema and both configs
  • Depends on: 2 · Requirements: R5.1
  • Test: T10 — make validate; T1 — P3
• [x] 4. Point the eight human-read templates at the skill
  • requirements, bugfix, design, testing-plan, tasks, pr-briefing, decision, capability
  • Depends on: 3 · Requirements: R2.1, R2.2
  • Test: T1 — P2, P3 (green)
• [x] 5. Wire the skill into the operating model
  • One bullet in skills/the-loop/SKILL.md; the three surveyed skills into externalTools
  • Depends on: 2 · Requirements: R1.4
  • Test: T10 — make validate
• [x] 6. Fold in the docs
  • docs/capabilities/writing-style.md + index, decision-061/062 + index, docs/config/harness-config.md, README.md, site nav
  • Depends on: 5 · Requirements: R1.4, R5.1
  • Test: T1 — P4; make lint
• [x] 7. Execute testing-plan.md and commit the evidence
  • Depends on: 4, 6 · Requirements: all
  • Test: T1, T8, T10, T11
• [x] 8. Remove length budgets (owner's review decision, PR #168)
  • Schema, both configs, the eight template pointers, the skill, the test (P1–P4), decision-061, and the requirements/design/testing-plan they came from
  • Depends on: 7 · Requirements: R2.1, R2.2, R2.3, R5.1, R5.2
  • Test: T1 — P3 asserts writingStyle.budgets stays absent; T10 — make validate

Dependency graph (DAG)

flowchart LR
  1["1 · test red"] --> 2["2 · skill"]
  2 --> 3["3 · schema + configs"]
  2 --> 5["5 · wire into the loop"]
  3 --> 4["4 · template pointers"]
  5 --> 6["6 · docs fold-in"]
  4 --> 7["7 · verify + evidence"]
  6 --> 7
  7 --> 8["8 · remove budgets<br/>(review decision)"]

Checkpoints

After task 1 (red recorded), after task 4 (make test green), after task 7 (make check clean, evidence committed) and after task 8 (make check re-run against the reduced contract). Then the review chain and the security-review gate. Risk tier 4 — completion waits for the named human approval on the PR.

Review comments

Appended by the-loop's record-feedback hook when a human gate approves with comments (issue-109).