Skip to content

chore(history): rehost archived m-tools design docs under .github/docs/history/#14

Closed
rafael5 wants to merge 1 commit into
mainfrom
chore/rehost-archived-m-tools-docs
Closed

chore(history): rehost archived m-tools design docs under .github/docs/history/#14
rafael5 wants to merge 1 commit into
mainfrom
chore/rehost-archived-m-tools-docs

Conversation

@rafael5
Copy link
Copy Markdown
Contributor

@rafael5 rafael5 commented May 11, 2026

Summary

  • Rehost the three load-bearing design docs from the now-archived m-dev-tools/m-tools repo as verbatim, frozen snapshots under .github/docs/history/.
  • Drop the tool:m-tools entry from profile/tools.json and retarget the three task_index.json history intents at new doc:m-dev-tools#<filename> typed IDs pointing at the in-org rehosts.
  • Update the profile/README.md table row + architecture diagram caption and reconcile the two AI-discoverability docs (table row in §3.4 marked resolved; non-goal docs(CONTRIBUTING): add org-level Layout conventions section #4 expanded to record the resolution; one illustrative typed ID corrected).

Resolves AI-discoverability-plan.md §8 non-goal #4: "m-tools is archived; either rehost its docs in .github/docs/history/ or drop them from tools.json. Agents care about the current shape." Chose rehost because the three preserved docs (gap analysis + Tier 1 strategy + remediation roadmap) are the design rationale behind m-cli's whole CLI shape; losing them to upstream pruning would erase the why. Cost: ~262 KB of frozen Markdown in .github.

Files

  • Newdocs/history/{README.md, m-tool-gap-analysis.md, m-tooling-tier1.md, gap-analysis-and-remediation-strategy.md}. Each rehost carries a single > Archived snapshot. banner after its H1 citing upstream commit 16fe3f7 (2026-04-27). One dangling sibling-link to implementation.md in gap-analysis-and-remediation-strategy.md was retargeted at the archived upstream URL (read-only on GitHub) rather than rehosting a fourth file — provenance noted in history/README.md.
  • Modifiedprofile/tools.json (drop tool:m-tools block + update first-principles bullet), profile/task_index.json (retarget 3 history intents to doc:m-dev-tools#<slug> typed IDs + new in-org URLs), profile/README.md (Historical-root row + diagram caption), docs/AI-discoverability-plan.md (§3.4 row + §8 non-goal docs(CONTRIBUTING): add org-level Layout conventions section #4), docs/AI-discoverability-guide.md (one example typed ID).

Test plan

  • python3 -m json.tool profile/tools.json — parses
  • python3 -m json.tool profile/task_index.json — parses
  • make check-docs-prose — green (.github/docs/history/ is .md-only)
  • make validate-catalog — green
  • Typed-ID grammar regex from tools.schema.json#/\$defs/typedID applied to every primary + see_also in task_index.json — all pass; three new IDs validated:
    • doc:m-dev-tools#gap-analysis-and-remediation-strategy
    • doc:m-dev-tools#m-tool-gap-analysis
    • doc:m-dev-tools#m-tooling-tier1
  • Grep audit: no remaining live m-tools/blob/main/docs/... URLs in the catalog or routing files; remaining m-tools mentions are intentional provenance citations only.

Not in scope

  • m-tools/docs/implementation.md and ydb-dev-tools-gap-analysis.md are deliberately not rehosted (history/README.md explains why). Both remain reachable in the archived m-tools repo for anyone who wants the deeper context.

🤖 Generated with Claude Code

…s/history/

The m-tools repo on GitHub is archived. Three design documents in it
are still load-bearing — they're the *why* behind the m-dev-tools
ecosystem's shape (the Go/Rust/Python toolchain analogy that drove
m-cli's CLI ergonomics, the Tier-1 strategy that scoped the first
m-cli release, and the phased remediation roadmap that produced
both m-cli and m-stdlib). Routing agents at a repo we no longer
maintain — and one that GitHub could prune at any time — is brittle.

Resolution per AI-discoverability-plan.md §8 non-goal #4 ("rehost
its docs in .github/docs/history/ or drop them from tools.json"):
rehost, not drop.

Changes:

- New: docs/history/ — verbatim copies of the three preserved docs,
  each with a single `> Archived snapshot.` banner after the H1
  citing the upstream commit (16fe3f7, 2026-04-27). README.md in
  the directory documents the provenance policy and the explicit
  exclusion of implementation.md + ydb-dev-tools-gap-analysis.md.
- profile/tools.json: dropped the tool:m-tools entry; updated the
  first-principles bullet to point at the in-org rehost path.
- profile/task_index.json: retargeted the three history-category
  intents to doc:m-dev-tools#<filename> typed IDs and the new in-org
  doc URLs.
- profile/README.md: replaced the "Historical root" repo row and
  the architecture-diagram caption to point at .github/docs/history/
  instead of the archived repo.
- docs/AI-discoverability-plan.md: marked the m-tools row in §3.4
  resolved; expanded non-goal #4 to record the resolution.
- docs/AI-discoverability-guide.md: updated the example typed ID
  in §11 from doc:m-tools#gap-analysis to
  doc:m-dev-tools#m-tool-gap-analysis to match the new scheme.

Validation:

- make check-docs-prose: green (only .md files under docs/history/).
- make validate-catalog: green.
- Typed-ID grammar regex (tools.schema.json $defs/typedID) applied
  to every primary + see_also in task_index.json: all pass.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@rafael5
Copy link
Copy Markdown
Contributor Author

rafael5 commented May 11, 2026

Superseded by #17 (rebuilt against post-P1-A/P1-D baseline; this PR's profile/tools.json was on the pre-P1-A shape and would have produced unresolvable merge conflicts). Same content; same intent.

@rafael5 rafael5 closed this May 11, 2026
@rafael5 rafael5 deleted the chore/rehost-archived-m-tools-docs branch May 11, 2026 03:42
rafael5 added a commit that referenced this pull request May 11, 2026
…s/history/ (#17)

The m-tools repo on GitHub is archived. Three design documents in it
are still load-bearing — they're the *why* behind the m-dev-tools
ecosystem's shape (Go/Rust/Python toolchain analogy that drove m-cli's
CLI ergonomics, Tier-1 strategy that scoped the first m-cli release,
phased remediation roadmap that produced both m-cli and m-stdlib).
Routing agents at a repo we no longer maintain — and one GitHub could
prune at any time — is brittle.

Resolution per AI-discoverability-plan.md §8 non-goal #4 ("rehost its
docs in .github/docs/history/ or drop them from tools.json"):
rehost, not drop.

Changes:

- docs/history/ — verbatim copies of three preserved docs:
  - gap-analysis-and-remediation-strategy.md
  - m-tool-gap-analysis.md
  - m-tooling-tier1.md
  Plus a README.md documenting the provenance policy and the explicit
  exclusion of implementation.md + ydb-dev-tools-gap-analysis.md.
  Each preserved doc carries a single `> Archived snapshot.` banner
  after the H1 citing upstream commit 16fe3f7 (2026-04-27).

- profile/tools.json: dropped the `m-tools` entry (was `status:
  archived`, no manifest by design); build-catalog's archived-carryover
  no longer triggers for it since the prior tools.json doesn't list it.

- profile/task_index.json: retargeted the three `history` category
  intents from `tool:m-tools` to typed-ID `doc:m-dev-tools#<filename>`
  and the new in-org doc URLs.

- profile/README.md, docs/AI-discoverability-plan.md,
  docs/AI-discoverability-guide.md: cross-references retargeted at the
  in-org paths.

Equivalent to the original chore/rehost-archived-m-tools-docs work
(PR #14), rebuilt on top of post-P1-A/P1-D state. PR #14 conflicted
heavily on profile/tools.json because it was authored against
pre-P1-A baseline.

Verification:
- make check-docs-prose      ✓ (docs/history/ holds only .md prose)
- make validate-catalog      ✓ (tools.json + task_index.json strict-validate)
- make catalog               ✓ (regen is byte-idempotent; m-tools stays gone)
- 26/26 pytest in profile/build/
- make phase0-smoke          PASS
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant