refactor(skills): retrofit the corpus to the ADR-0020 context contract (#129)

Retrofits all 39 skills to ADR-0020's description/body context contract, then fixes what six rounds of independent review found in that retrofit — including four ways the hot gate itself failed open.

Closes #99, #107, #108, #110, #111, #114, #115, #120.

## The retrofit (waves 1-5)

| | Start | Now |
|---|---|---|
| Description FAILs (>400 chars) | 26 | **0** |
| Body FAILs (>900 words, body-only) | 9 | **0** |
| Dangling routing targets | 2 | **0** |
| `Kyberforge.CompositionNote` | 10 | **0** |
| Preload tax | 21,005 chars | **~10,500** |

Under the 12,000-char success criterion. Per-wave detail is on #99.

## The review fixes

**The gate failed open four ways, three of them found after the retrofit shipped.** An unrecognised follower token made a dangling target vanish. A skill directory with no `SKILL.md` resolved as a valid target, so a commit could be green locally and red in a fresh clone — three existing fixtures were relying on that, one of which made the install-leak A/B pass vacuously. Then the free-standing `/name` sweep turned out to be gated on the sentence carrying a boundary marker, so route notation in any other sentence was invisible — not an ERROR, not a SUGGESTION, not an INFO — which left the documented "`/name` always blocks" promise false from a second direction. All four fixed and pinned.

**Two checks were silently not running.** `validate-provenance.sh` checks 7-8 were dead across nine skills. Waking them exposed a deeper problem: they assume `Research doc:` names a source index, but 30 of 121 entries point at topic content documents, so every new check-7 INFO was a false positive and check 8 was saved from a false-FAIL flood only by an *unannounced* skip. Checks 7/8 are now scoped to source indexes and every skip announces itself (#121).

**The retrofit's own anti-goal, four times.** ADR-0020 warns that a blunt gate gets satisfied by deleting content rather than relocating it. `diagnose` and `skill-audit` relocated prose and then read it unconditionally; `prototype` and `vale-config` deleted rules outright that survived nowhere. All four addressed.

## Verification

- `bash tests/run-tests.sh --strict` — 24 suites, 0 skipped, 0 failed
- `bash tests/run-bats.sh` — 325 tests, 0 failures
- `pre-commit run --all-files` — 17/17
- `pre-commit run --hook-stage pre-push --all-files` — 16/16, with `apm marketplace check` and `apm pack --check-clean` run against the remote, not skipped
- `scripts/skill-size-check.sh` over all 39 skills — rc 0, 0 ERROR/FAIL, SUGGESTION-only
- Preload tax measured at **10,498 chars**, max description 390 — both inside budget
- Every new test proven non-vacuous by a deliberate mutation of the behaviour it covers

**Per-commit sync, stated accurately:** the ten commits from the latest review round each pass `check-plugin-content-sync` in isolation, verified by checking each out in a detached worktree with a clean between. The earlier gitea window (`dfacf05..bedbd1d`, nine commits) does **not** — its mirror was regenerated in one batch at `bbc7300`. An earlier revision of this description claimed the property held for every commit; it does not, and a bisect through that window lands on a red commit. **Squash-merge** to collapse it, or accept that this range is not bisectable.

## Version bump

Six plugins and the catalog take a **patch**, not a minor. The branch is **89 commits — 40 `fix` / 30 `refactor` / 12 `docs` / 5 `chore` / 2 `test` — zero `feat`, zero `!`, zero `BREAKING CHANGE`** — and adds no skill, agent, command or hook. (Two earlier revisions of this section cited a stale histogram, most recently 78 commits; the figures above are measured at HEAD.) Both rules this repo ships (`forge/references/version-bump.md`, landing in this PR, and `git-commits/references/conventional-commits-spec.md`) make that a patch, and the catalog set is unchanged at 7 entries.

Not settled by that: four published files were removed from the installed tree, three moved, and `caveman` gained `disable-model-invocation`, retiring its old triggers. Under a strict reading those are major-class and currently ship under `refactor:` with no marker. Whether the deployed skill surface is a public contract is written down nowhere — worth deciding, but it outlives this PR.

## Deliberately not in scope

#112 (cherry-pick ownership, now resolved in favour of `git-commits`), #113 (`rtk git` normalisation), #116 (research fan-out), #101 (audit-skill merge), #122 (non-spec skill-root files), #123 (no PRD producer) stay open. #117 is the one worth reading: the contract's remedy is to move prose into `references/`, which is exactly where neither the size gate nor Vale looks — and the blind spot is wider than #117 currently records, since there is no root `.vale.ini` at all, so every ADR, `CONTEXT.md` and `README.md` is unlinted too.

That blind spot let this branch carry two `level: error` `Kyberforge.SentenceOpenerThereIs` violations into `references/` files it created — `provider-adapter-author/references/provider-matrix.md:31` and `agent-audit/references/finding-criteria.md:95`. Both are reworded in `afadaae`, confirmed by routing each file through the audit's own `vale-wrap.sh` (1 error each before, 0 after). Five further occurrences sit in `references/` files already on `main`; those are the pre-existing corpus and stay with #117, which is the real fix.

Also unfixed and not this PR's: `apm install` appends a duplicate `SessionStart` entry to `.claude/settings.json`, so a fresh clone cannot get pre-push green without an edit AGENTS.md warns against. Reproduces identically on `main`.

Co-authored-by: Defame1297 <gitea@rkdr.net>
Reviewed-on: https://git.dev.rkdr.net/Defame1297/holocron/pulls/129
Co-authored-by: Claude Code AI - Gitea MCP <claude@noreply.git.dev.rkdr.net>
Co-committed-by: Claude Code AI - Gitea MCP <claude@noreply.git.dev.rkdr.net>

plugins/git/skills/git-commits/references/rewrite-history.md

@@ -0,0 +1,76 @@
+---
+source_keys:
+  - org-commit-conventions
+  - context7-git-htmldocs
+---
+
+# Rewriting existing commits
+
+Every flow on this page rewrites history. None of them runs before the caller has explicitly approved it, and none is followed by a force-push to `main`/`master` — refuse that and explain why instead.
+
+## Amend the last commit
+
+1. Stage the new changes, or the changes that undo something.
+2. Run `rtk git commit --amend`, adding `--no-edit` when the message stays as it is.
+3. If the message should change, show the current one and prompt for the replacement.
+4. The branch has now diverged from its remote. Amending is safe only on a branch nobody else has based work on; on `main`/`master`, refuse the force-push and explain, rather than warning and proceeding.
+
+## Fold a commit into an earlier one (autosquash — preferred)
+
+Prefer this whenever a commit is written to be folded, because git does the marking:
+
+1. `rtk git commit --fixup=<commit>` keeps the target's message; `rtk git commit --squash=<commit>` lets you edit the combined message later. Both prefix the message with `fixup!`/`squash!` and name the target commit.
+2. Get explicit approval — the rebase still rewrites history.
+3. Run `rtk git rebase -i --autosquash HEAD~N`. Git pre-fills the todo list with the tagged commits already reordered against their targets; save it unchanged to apply.
+
+**`-i` is not optional here.** On Git 2.39.5, `rtk git rebase --autosquash HEAD~N` without `-i` prints `Successfully rebased and updated refs/heads/<branch>.` and exits 0 while leaving the `fixup!` commit in place at its original SHA — `--autosquash` is honoured only by the interactive machinery, and the false success is the trap: the fold is reported as done, and the surviving `fixup!` subject then fails the Conventional Commits `commit-msg` hook. Later Git versions taught the non-interactive rebase to honour the flag, but `-i --autosquash` is correct on every version, so always write that.
+
+## Squash by hand (interactive rebase)
+
+Use this when the commits were not tagged at commit time. **Interactive rebase has no undo once `rebase -i` starts — `git reflog` is the recovery path.**
+
+1. Identify the commits to squash — typically the last N on the current branch.
+2. Get explicit approval.
+3. Run `rtk git rebase -i HEAD~N`, marking the older commits `squash` to keep their messages for editing, or `fixup` to discard them.
+4. Compose the combined message when the rebase stops to ask. For a non-trivial combined message, follow the structure in `references/commit-template.md`.
+
+## When a rebase halts on a conflict
+
+Offer conflict resolution or `rtk git rebase --abort`. Do not resolve conflicts automatically without confirmation.
+
+## Rebase the branch onto a new base
+
+Replays this branch's commits on top of another branch's tip — bringing a feature branch up to
+date without a merge commit.
+
+1. Confirm nothing being replayed has been pushed anywhere someone else has based work on. A rebase
+   gives every replayed commit a new SHA, which breaks any clone that already has the old ones.
+2. Get explicit approval — this rewrites history like every other flow on this page.
+3. `rtk git fetch origin` first, so `<newbase>` is the real tip rather than a stale local copy.
+4. `rtk git rebase <newbase>` — for example `rtk git rebase main`. Use
+   `rtk git rebase --onto <newbase> <upstream> <branch>` to replay only the commits after
+   `<upstream>`, which is how a branch started from the wrong base gets moved.
+5. The branch has now diverged from its remote. It needs
+   `--force-with-lease --force-if-includes` to push, never a bare `--force`, and never on
+   `main`/`master` — refuse that and explain.
+
+## Move the branch pointer back (`git reset`)
+
+`reset` moves the current branch to another commit. The mode decides what survives:
+
+| Mode | Branch pointer | Index | Working tree |
+|---|---|---|---|
+| `--soft` | moves | untouched — the changes stay staged | untouched |
+| `--mixed` (default) | moves | reset — the changes become unstaged | untouched |
+| `--hard` | moves | reset | **overwritten; uncommitted work is destroyed** |
+
+- "Undo my last commit but keep the changes" is `rtk git reset --soft HEAD~1`. That is the default
+  answer to the request; reach for anything else only when the caller asked for it.
+- `rtk git reset --mixed HEAD~1` when the staging should be redone from scratch too.
+- `rtk git reset --hard <ref>` is gated like a force-push: state exactly which uncommitted changes
+  will be discarded, get approval for that specific reset, and offer `rtk git stash push -u` first.
+  The commits it drops stay reachable through `git reflog`; the uncommitted edits never entered git
+  at all and nothing recovers them.
+
+Reset does not rewrite the commits it leaves behind, so no force-push is needed unless the branch
+was already pushed at the newer commit.