Tasks: the learnings tree is a configured location, and it defaults into docs/

Phase 3 of 3. A DAG, not a list: tasks with the same dependencies can run in any order. Each _Test:_ names a row of testing-plan.md.

graph LR
  T1["1 · schema key"] --> T2["2 · template + packaged default"]
  T1 --> T3["3 · this repo's config"]
  T1 --> T6["6 · docs site"]
  T4["4 · git mv the tree"] --> T5["5 · manifest"]
  T4 --> T6
  T1 --> T7["7 · skill + reference"]
  T1 --> T8["8 · commands"]
  T4 --> T7
  T4 --> T8
  T1 --> T9["9 · decision-082"]
  T7 --> T10["10 · capability doc"]
  T2 --> T11["11 · verification"]
  T3 --> T11
  T5 --> T11
  T6 --> T11
  T8 --> T11
  T9 --> T11
  T10 --> T11

Tasks

• [x] 1. Add workflow.learningsDir to the schema. String, default docs/learnings, described in the register specDir/capabilitiesDir use: the three paths it governs and the repo-relative rule. Re-point the two selfImprovement descriptions that name the overflow path at <workflow.learningsDir>, and extend the workflow onboarding group's explain so the guided init mentions learnings alongside specs and capability docs. Requirements: R1.1–R1.4, R2.1 · Test: T1

• [x] 2. State the default in the template and the packaged default. Add learningsDir: docs/learnings under workflow in skills/the-loop/templates/harness-config.yaml with the inline comment style its siblings carry, then copy the file to cli/the_loop/harness-config.default.yaml so the two stay byte-identical. Requirements: R2.1, R2.2 · Test: T1, T2, T3

• [x] 3. Declare it in this repository's own config. learningsDir: docs/learnings in .the-loop/harness-config.yaml, stated rather than defaulted. Requirements: R3.3 · Test: T1, T3

• [x] 4. Move the tree. git mv learnings docs/learnings, then fix the relative links inside the moved files (topics/README.md points at the skill reference; the index points at its records). Requirements: R3.1, R3.2, NFR-4, NFR-5 · Test: T8, T10, T18

• [x] 5. Update the manifest. The three knowledge entries become docs/learnings/learnings.md, docs/learnings/learning-<nnn>.md and docs/learnings/topics/<category>.md. Requirements: R4.2 · Test: T3, T8

• [x] 6. Update the documentation site. The workflow row of docs/config/harness-config.md names the new key; the layout trees in docs/guide/how-it-works.md and docs/architecture/architecture.md show the new location. Requirements: R4.3 · Test: T5, T7, T8

• [x] 7. Update the skill and the automation reference. SKILL.md §Knowledge the loop maintains and reference/automation.md §Self-improvement name <learningsDir>/… and state the key and its default once. Requirements: R4.1 · Test: T7, T8

• [x] 8. Update the commands. /the-loop:init scaffolds the index under the configured directory; /the-loop:work-on and /the-loop:execute-tasks write there; /the-loop:upgrade-the-loop gains the relocation paragraph — present both outcomes, move nothing without confirmation. Requirements: R4.1, R5.1, R5.2 · Test: T8, T9

• [x] 9. Record decision-082. The placement question (workflow over selfImprovement) and the default relocation, with the alternatives and their costs; add the index row. Requirements: R4.3 · Test: T7

• [x] 10. Fold into the capability doc. docs/capabilities/spec-workflow.md — the doc that owns the workflow.* directory keys — gains learningsDir and a history row. Requirements: R4.3 · Test: T5, T7

• [x] 11. Execute the testing plan. Run every activity, record command, outcome and evidence in evidence/verification.md, tick each row only once it has actually run, and complete testing-plan.md as the record. Requirements: all · Test: T1–T10, T15, T18