docs: consolidate platform/docs/ via trueup pipeline #6

Merged
pappi merged 1 commit from agent/docs-trueup-2026-05-09 into main 2026-05-10 10:32:47 +00:00
Owner

What

Trueup pass on platform/docs/. Reduces the tree from 1,360 files
to 69
by aggregating thematic clusters into consolidated documents,
relocating archive content to a new platform/archive/ sibling, and
deleting source files whose content is now in the consolidated docs.

Tooling: scripts + rules + prompts at /home/mreddy/hanuman/temp/
(operator workstation only — not committed; this PR's contents are
the output, not the pipeline). The 5-stage pipeline (inventory →
cluster → aggregate → manual review → execute) was run end-to-end
under the Max plan via subagents (no API spend).

15 consolidated documents written:

Path Sources Type
docs/architecture/workspace-model.md 3 related-set
docs/architecture/architecture-extracts.md 2 version-pair
docs/archive/handoff-artifacts/machine-protocol-handoff.md 11 related-set
docs/archive/historical-reports/stage-reports.md 4 related-set
docs/archive/historical-reports/hanuman-ops-roadmap.md 21 related-set
docs/archive/deferred-decisions/rebaseline.md 5 related-set
archive/snapshots/automation-context.md 16 snapshot
archive/snapshots/cp-sonar-poc-bundle-summary.md 210 code-archive
archive/snapshots/cp-sonar-poc-bundle.tar.gz 1 preserved tarball
archive/historical-reports/contract-migration-prep.md 2 snapshot
archive/historical-reports/contract-reconciliation.md 4 snapshot
archive/workflow-history/task-history.md 23 workflow-outputs
archive/workflow-history/public-repo-reviews-summary.md 764 workflow-outputs
archive/workflow-history/hanuman-ops-roadmap-history.md 6 workflow-outputs
archive/code/hsdk-scripts-summary.md 72 code-archive
archive/code/hsdk-tools-summary.md 34 code-archive

Per-file relocations (no aggregation):

  • 40 source-code files (code-files-in-docs cluster) → archive/code/<original-path>
  • 80 misc archive markdown / txt / log / jsonl → archive/deprecated/<original-path> (directory structure preserved to avoid basename collisions on README.md etc.)
  • 316 originals preserved alongside their consolidated summary at archive/code/{hsdk-scripts,hsdk-tools,cp-sonar-poc-bundle}/<inner-path>

Datestamps stripped from filenames per workspace convention; the
captured-on dates are preserved inside each document body.

33 preserve-as-is files left untouched: the strategic doc set
(operating-thesis, strategy, decisions-log, technical-architecture,
architecture-summary, partnership-and-channel-map,
next-90-days-execution-plan, recalibration-log), the
architecture/components/ quartet, the architecture/foundation/
trio, the PEX-0001 ADR, governance/ledger.json, all of
policies/, templates/, contracts/, ops/, pattern-library/,
handovers/.

Why

platform/docs/ had accumulated organically over months. 263 source
code files, 687 JSON workflow outputs, 35+ README.md files in archive
sub-trees, multiple version-pair documents (v1 and v2 sitting
alongside each other), 764 sequential workflow-run dirs from one
day's bring-up against apache/commons-lang. Without consolidation,
finding current architectural truth required wading through
historical strata; the foundation work landing today (PR #3) needs a
clean substrate going forward.

The trueup pass:

  • Establishes platform/archive/ as the home for non-doc archive
    content (code, snapshots, workflow history) — keeping docs/ for
    documents only.
  • Converts 13 thematic clusters of related files into 15 coherent
    consolidated documents that read top-to-bottom.
  • Removes the source files whose substantive content is fully
    preserved in those consolidated docs.

Subagents flagged a handful of fidelity items worth post-merge
attention (recorded inline in the consolidated docs):

  • architecture-extracts.md flags v1↔v2 contradiction on the
    learning-loop liveness claim (preserves both characterizations).
  • machine-protocol-handoff.md flags F8-vs-F21 sequencing
    disagreement between two snapshots.
  • task-history.md discovered the 23 task files were not a
    temporal series but a single archival snapshot — reframed
    accordingly.

How verified

  • Pipeline script outputs reviewed at each stage; cluster plan
    inspected before aggregation; 15 aggregated docs spot-checked.
  • All 1,313 planned actions recorded in execute-log.json (not
    committed). Action breakdown: 15 copy-consolidated, 861
    delete-aggregated, 437 moves, 0 destination collisions.
  • Conforms to B3 git policy: branch agent/docs-trueup-2026-05-09
    matches the regex; commit follows Conventional Commits; this PR
    body has the three required sections.
  • Recovery path: git revert HEAD restores the entire pre-trueup
    tree state in one operation. Fine-grained recovery (specific
    documents only) requires git checkout HEAD~1 -- <path> for the
    relevant source files.

Refs: B3 (docs/wip/B3-git-policy.md), foundation v0
(docs/architecture/foundation/).

## What Trueup pass on `platform/docs/`. Reduces the tree from **1,360 files to 69** by aggregating thematic clusters into consolidated documents, relocating archive content to a new `platform/archive/` sibling, and deleting source files whose content is now in the consolidated docs. **Tooling:** scripts + rules + prompts at `/home/mreddy/hanuman/temp/` (operator workstation only — not committed; this PR's contents are the output, not the pipeline). The 5-stage pipeline (inventory → cluster → aggregate → manual review → execute) was run end-to-end under the Max plan via subagents (no API spend). **15 consolidated documents written:** | Path | Sources | Type | |---|--:|---| | `docs/architecture/workspace-model.md` | 3 | related-set | | `docs/architecture/architecture-extracts.md` | 2 | version-pair | | `docs/archive/handoff-artifacts/machine-protocol-handoff.md` | 11 | related-set | | `docs/archive/historical-reports/stage-reports.md` | 4 | related-set | | `docs/archive/historical-reports/hanuman-ops-roadmap.md` | 21 | related-set | | `docs/archive/deferred-decisions/rebaseline.md` | 5 | related-set | | `archive/snapshots/automation-context.md` | 16 | snapshot | | `archive/snapshots/cp-sonar-poc-bundle-summary.md` | 210 | code-archive | | `archive/snapshots/cp-sonar-poc-bundle.tar.gz` | 1 | preserved tarball | | `archive/historical-reports/contract-migration-prep.md` | 2 | snapshot | | `archive/historical-reports/contract-reconciliation.md` | 4 | snapshot | | `archive/workflow-history/task-history.md` | 23 | workflow-outputs | | `archive/workflow-history/public-repo-reviews-summary.md` | 764 | workflow-outputs | | `archive/workflow-history/hanuman-ops-roadmap-history.md` | 6 | workflow-outputs | | `archive/code/hsdk-scripts-summary.md` | 72 | code-archive | | `archive/code/hsdk-tools-summary.md` | 34 | code-archive | **Per-file relocations (no aggregation):** - 40 source-code files (`code-files-in-docs` cluster) → `archive/code/<original-path>` - 80 misc archive markdown / txt / log / jsonl → `archive/deprecated/<original-path>` (directory structure preserved to avoid basename collisions on README.md etc.) - 316 originals preserved alongside their consolidated summary at `archive/code/{hsdk-scripts,hsdk-tools,cp-sonar-poc-bundle}/<inner-path>` **Datestamps stripped from filenames** per workspace convention; the captured-on dates are preserved inside each document body. **33 preserve-as-is files left untouched:** the strategic doc set (operating-thesis, strategy, decisions-log, technical-architecture, architecture-summary, partnership-and-channel-map, next-90-days-execution-plan, recalibration-log), the `architecture/components/` quartet, the `architecture/foundation/` trio, the PEX-0001 ADR, `governance/ledger.json`, all of `policies/`, `templates/`, `contracts/`, `ops/`, `pattern-library/`, `handovers/`. ## Why `platform/docs/` had accumulated organically over months. 263 source code files, 687 JSON workflow outputs, 35+ README.md files in archive sub-trees, multiple version-pair documents (v1 and v2 sitting alongside each other), 764 sequential workflow-run dirs from one day's bring-up against `apache/commons-lang`. Without consolidation, finding current architectural truth required wading through historical strata; the foundation work landing today (PR #3) needs a clean substrate going forward. The trueup pass: - Establishes `platform/archive/` as the home for non-doc archive content (code, snapshots, workflow history) — keeping `docs/` for documents only. - Converts 13 thematic clusters of related files into 15 coherent consolidated documents that read top-to-bottom. - Removes the source files whose substantive content is fully preserved in those consolidated docs. Subagents flagged a handful of fidelity items worth post-merge attention (recorded inline in the consolidated docs): - `architecture-extracts.md` flags v1↔v2 contradiction on the learning-loop liveness claim (preserves both characterizations). - `machine-protocol-handoff.md` flags F8-vs-F21 sequencing disagreement between two snapshots. - `task-history.md` discovered the 23 task files were not a temporal series but a single archival snapshot — reframed accordingly. ## How verified - Pipeline script outputs reviewed at each stage; cluster plan inspected before aggregation; 15 aggregated docs spot-checked. - All 1,313 planned actions recorded in execute-log.json (not committed). Action breakdown: 15 copy-consolidated, 861 delete-aggregated, 437 moves, 0 destination collisions. - Conforms to B3 git policy: branch `agent/docs-trueup-2026-05-09` matches the regex; commit follows Conventional Commits; this PR body has the three required sections. - Recovery path: `git revert HEAD` restores the entire pre-trueup tree state in one operation. Fine-grained recovery (specific documents only) requires `git checkout HEAD~1 -- <path>` for the relevant source files. --- Refs: B3 (`docs/wip/B3-git-policy.md`), foundation v0 (`docs/architecture/foundation/`).
docs: consolidate platform/docs/ via trueup pipeline
All checks were successful
Branch policy / branch-policy (pull_request) Successful in 26s
8f69ca5eb4
Reduces platform/docs/ from 1,360 files to 69 by:

- Aggregating 13 thematic clusters into 15 consolidated documents
  (related-set, version-pair, snapshot, workflow-outputs, code-archive
  types per the rules in temp/rules/cluster-rules.json) using the
  prompts at temp/prompts/aggregate-*.txt.
- Relocating 446 files (source code + binary + ad-hoc archive content)
  to a new platform/archive/ tree (sibling of docs/) split into
  code/, deprecated/, snapshots/, historical-reports/, workflow-history/.
- Deleting 861 source files whose substantive content is now captured
  in the corresponding consolidated documents (recoverable from git
  history if any consolidation lost fidelity).

New consolidated documents at platform/docs/:
- architecture/workspace-model.md (3 sources)
- architecture/architecture-extracts.md (2 sources, version-pair)
- archive/handoff-artifacts/machine-protocol-handoff.md (11 sources)
- archive/historical-reports/stage-reports.md (4 sources)
- archive/historical-reports/hanuman-ops-roadmap.md (21 sources)
- archive/deferred-decisions/rebaseline.md (5 sources)

New consolidated documents at platform/archive/:
- snapshots/automation-context.md (16 sources)
- snapshots/cp-sonar-poc-bundle-summary.md (210 sources, code-archive)
- snapshots/cp-sonar-poc-bundle.tar.gz (preserved tarball move)
- historical-reports/contract-migration-prep.md (2 sources)
- historical-reports/contract-reconciliation.md (4 sources)
- workflow-history/task-history.md (23 sources)
- workflow-history/public-repo-reviews-summary.md (764 sources)
- workflow-history/hanuman-ops-roadmap-history.md (6 sources)
- code/hsdk-scripts-summary.md (72 sources, code-archive)
- code/hsdk-tools-summary.md (34 sources, code-archive)

Datestamps removed from filenames per workspace convention; the
captured-on dates are preserved inside each document body.

Preserved as-is (33 files): operating-thesis.md, strategy.md,
decisions-log.md, technical-architecture.md, architecture-summary.md,
partnership-and-channel-map.md, next-90-days-execution-plan.md,
recalibration-log.md, architecture/components/, architecture/foundation/,
ADRs, governance/ledger.json, policies/, templates/, contracts/, ops/,
pattern-library/, handovers/.

Pipeline scripts and rules in /home/mreddy/hanuman/temp/ (not
committed — operator workstation only). Recovery: git revert HEAD,
or git reset --hard to restore prior tree state.
pappi merged commit 5a3f319b85 into main 2026-05-10 10:32:47 +00:00
Sign in to join this conversation.
No reviewers
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
hanuman/platform!6
No description provided.