diff --git a/AGENTS.md b/AGENTS.md
index 8a520a3..3aeb20d 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -28,8 +28,8 @@ Fall back to raw shell only when no skill covers it.
 ## Session rules
 
 - **Do not add repo-owned keys to `.claude/settings.json`.** apm treats it as its own deployed artifact and `apm audit --ci` replays the install and diffs, so anything apm would not have written is permanent drift that fails the `apm-audit-ci` pre-push hook. A hook you want here is authored in `plugins/<name>/.apm/hooks/` and deployed by apm, never hand-written into that file. The `SessionStart` entry already in it is exactly that: kyberforge authors it in `plugins/kyberforge/.apm/hooks/hooks.json` and apm merges it in, so it is apm's own output, it is what the replay expects, and it belongs in the commit — do not strip it (ADR-0019). Machine-specific settings go in the gitignored `.claude/settings.local.json`; shared enforcement goes in `.pre-commit-config.yaml`.
-- **`apm.lock.yaml` turning up modified is expected, not a bug.** kyberforge's `SessionStart` hook keeps the install current on launch and rewrites the lock in the process (ADR-0019). Commit or discard it deliberately.
-- **A `.apm/` edit is not live in this session until it is pushed.** The six dependencies resolve from the holocron remote, unpinned against the default branch. `apm install` deploys from the lock; `apm update` is what re-resolves refs.
+- **`apm.lock.yaml` turning up modified is expected, not a bug.** kyberforge's `SessionStart` hook keeps the install current on launch and rewrites the lock in the process (ADR-0019). On `main`, commit or discard it deliberately; on a feature branch, discard it (`git checkout -- apm.lock.yaml`, then `apm install`) — the refresh resolved against `main`, not the branch.
+- **A `.apm/` edit is not live until it is on the remote's `main`.** The six dependencies resolve from the holocron remote, unpinned against the default branch, so pushing a feature branch does not deploy it (ADR-0019). `apm install` deploys from the lock; `apm update` is what re-resolves refs.
 - **No pre-push hook needs the network.** Root `apm.yml`'s marketplace has no remote package entries, so every hook resolves locally.
 - **This repo and Gitea are the only source of truth.** All project state, decisions, and working conventions live here. Do not use an external memory system for this project — cached state diverges from the repo and you get a split brain. Before answering any design or architecture question, check `docs/adr/` for an existing decision.
 
diff --git a/README.md b/README.md
index 348aa9e..cf207d8 100644
--- a/README.md
+++ b/README.md
@@ -31,7 +31,7 @@ Install all of these before setting up. Each one is a hard dependency of a git h
 | Tool | Why | Install |
 | --- | --- | --- |
 | `apm` CLI | Two pre-push hooks shell out to it (`apm-audit-ci` and `apm-pack-check-clean`) | The `apm-install` skill, or `curl -sSL https://aka.ms/apm-unix \| sh`. Verify with `apm --version` |
-| `python3` + PyYAML | Required by `scripts/skill-size-check.sh` (the `skill-size-check` pre-commit hook), which reads folded YAML frontmatter | `python3` is usually present — pre-commit is itself a Python application. `pip install pyyaml` if the hook reports PyYAML missing |
+| `python3` + PyYAML | Required by `scripts/skill-size-check.sh` (the `skill-size-check` pre-commit hook) and `scripts/check-skill-version-bump.sh` (the `check-skill-version-bump` pre-push hook), which both parse YAML frontmatter | `python3` is usually present — pre-commit is itself a Python application. `pip install pyyaml` if the hook reports PyYAML missing |
 | `vale` | Required by the `vale-audit-prefilter-skill` / `-agent` pre-commit hooks, and by the `test-vale-wrap.sh` / `test-vale-hooks-consumer.sh` suites that `run-tests --strict` runs at pre-push | `brew install vale` (macOS), `snap install vale` (Linux), `choco install vale` (Windows), or https://vale.sh/docs/vale-cli/installation/ |
 | `claude` CLI | Required by the `validate-marketplace` pre-push hook | Claude Code |
 
@@ -58,7 +58,7 @@ pre-commit install -t pre-commit -t commit-msg -t pre-push
 
 ## Keeping the install current
 
-The six dependencies in root `apm.yml` are unpinned against the default branch, so deployed skills go stale whenever anyone merges. kyberforge's `SessionStart` hook keeps the install current automatically on launch, rewriting `apm.lock.yaml` in the process — an unexplained modification to it after opening a session is expected, not a bug; commit or discard it deliberately. Mechanism and rationale: `docs/adr/0019-session-start-hook-keeps-the-apm-install-current.md`.
+The six dependencies in root `apm.yml` are unpinned against the default branch, so deployed skills go stale whenever anyone merges. kyberforge's `SessionStart` hook keeps the install current automatically on launch, rewriting `apm.lock.yaml` in the process — an unexplained modification to it after opening a session is expected, not a bug. On `main`, commit or discard it deliberately; on a feature branch, discard it (`git checkout -- apm.lock.yaml`, then `apm install`), because the refresh resolves against `main` and the lock now records `main`'s commit, not the branch's. Mechanism and rationale: `docs/adr/0019-session-start-hook-keeps-the-apm-install-current.md`.
 
 Note the difference between the two commands:
 
diff --git a/docs/adr/0019-session-start-hook-keeps-the-apm-install-current.md b/docs/adr/0019-session-start-hook-keeps-the-apm-install-current.md
index 34e44d1..f9352d3 100644
--- a/docs/adr/0019-session-start-hook-keeps-the-apm-install-current.md
+++ b/docs/adr/0019-session-start-hook-keeps-the-apm-install-current.md
@@ -115,9 +115,8 @@ A test pins the reference.
 > emptied it. `tests/test-apm-current-hook.sh` still pins the literal string.
 
 **Session startup gets slower when the install is stale.** Measured: ~0.7 s for the `apm outdated`
-check when everything is current, ~10.4 s when six packages are behind and the refresh runs
-(re-measured 2026-09-16: ~24–26 s for the same six-behind refresh, warm, on a LAN remote — still
-well inside the budget below). The hook declares `timeout: 380` to cover a cold multi-package fetch. That number is not free-standing:
+check when everything is current, ~10.4 s when six packages are behind and the refresh runs. The
+hook declares `timeout: 380` to cover a cold multi-package fetch. That number is not free-standing:
 the script imposes its own `timeout 60` on `apm outdated` and `timeout 300` on `apm update`, so the
 host-side timeout has to exceed their sum or the host kills the hook mid-update and leaves
 `.claude/skills/` half-deployed with no notice emitted. An earlier revision declared `320`, which
@@ -125,6 +124,10 @@ was below the 360 s the script can legitimately take. A test asserts the invaria
 literal — it parses every `timeout N` out of the script, sums them, and requires the `hooks.json`
 value to be larger — so changing either side without the other fails the suite.
 
+> **Amendment (2026-09-16) — the refresh is slower than first measured, still inside the budget.**
+> Re-measured: ~24–26 s for the same six-behind refresh, warm, on a LAN remote — well inside the
+> 380 s above.
+
 **Reading a human-readable CLI for a control decision cost a silent failure, again.** `apm outdated`
 has no `--json` or other machine-readable flag (confirmed against 0.28.0), so the hook must match
 its prose. The first attempt matched `outdated dependencies found` — plural only. apm emits
@@ -146,20 +149,21 @@ to plural-only fails it.
 deploy until this change is merged and `apm update` has run once against the new default branch.
 Until then the repo has the mechanism in source and not in effect.
 
-**On a feature branch, the refresh installs `main`, not the branch.** Added 2026-09-16, after it
-happened. The dependencies resolve against the remote default branch, so a session opened on a
-branch that changes `plugins/` loads `main`'s content, refreshed or not — the branch's own edits
-are never live until merged. Two visible effects follow. Content the branch *removes* comes back in
-the deployed install: on `docs/simplification-audit` a refresh redeployed `main`'s `skill-audit`
-and `agent-audit` over the branch's merged `factory-audit`, and re-materialised `main`'s
-`plugins/bin/.mcp.json` into `apm_modules/`, so the gitignored root `.mcp.json` regained the
-`obsidian` server the branch deleted — invisible to `git status`. And the rewritten
-`apm.lock.yaml` records `main`'s commit, so on a branch it should be discarded
-(`git checkout -- apm.lock.yaml`, then `apm install` to bring the deployed tree back in line with
-the lock, or `apm pack --check-clean` refuses to run), not committed. Skipping the refresh off the
-default branch was considered and rejected: it would not make the branch live, only freeze the
-session on an older `main` — the silent staleness this ADR exists to prevent. Both effects end when
-the branch merges; apm removes a server that is no longer declared on its next update.
+> **Amendment (2026-09-16) — on a feature branch, the refresh installs `main`, not the branch.**
+> Recorded after it happened. The dependencies resolve against the remote default branch, so a
+> session opened on a branch that changes `plugins/` loads `main`'s content, refreshed or not — a
+> branch's own `.apm/` edits are live only once they are on the remote's `main`. Two visible
+> effects follow. Content the branch *removes* comes back in the deployed install: on
+> `docs/simplification-audit` a refresh redeployed `main`'s `skill-audit` and `agent-audit` over
+> the branch's merged `factory-audit`, and re-materialised `main`'s `plugins/bin/.mcp.json` into
+> `apm_modules/`, so the gitignored root `.mcp.json` regained the `obsidian` server the branch
+> deleted — invisible to `git status`. And the rewritten `apm.lock.yaml` records `main`'s commit,
+> so on a branch it should be discarded (`git checkout -- apm.lock.yaml`, then `apm install` to
+> bring the deployed tree back in line with the lock, or `apm pack --check-clean` refuses to run),
+> not committed. Skipping the refresh off the default branch was considered and rejected: it would
+> not make the branch live, only freeze the session on an older `main` — the silent staleness this
+> ADR exists to prevent. Both effects end when the branch merges; apm removes a server that is no
+> longer declared on its next update.
 
 **`.claude/settings.json` stops being `{"hooks": {}}`.** apm merges the hook into it and tracks
 ownership in a `.claude/apm-hooks.json` sidecar, with the script copied to
diff --git a/docs/adr/0022-skill-metadata-version-is-mandatory.md b/docs/adr/0022-skill-metadata-version-is-mandatory.md
index 9ba050a..b507ca4 100644
--- a/docs/adr/0022-skill-metadata-version-is-mandatory.md
+++ b/docs/adr/0022-skill-metadata-version-is-mandatory.md
@@ -77,38 +77,49 @@ presence is no longer in question. `.pre-commit-config.yaml`'s `skill-frontmatte
 to require the field, closing the gap #113 and #118 both named in the same audit pass: a stated rule
 with nothing enforcing it drifts the same way an unstated one does.
 
-## Amendment (2026-09-16) — the bump is enforced at push, not only required to exist
+## Amendment (2026-09-16): the bump is enforced at push, not only required to exist
 
-**Context.** Making the field mandatory did not make it move. The only thing that bumped it was
-`skill-author` Step 4, so every hand edit and every trim pass skipped the bump: on
-`docs/simplification-audit`, 17 of the 40 skill directories that changed against `main` carried
-the same `metadata.version` as `main`, and `gitea` alone sat at six different values. Both
-validators checked presence and semver shape, never movement, so the field could not answer the
-question this ADR gives it — "did this change since I last read it". (Simplification audit
-finding 33.)
+Making the field mandatory did not make it move. The only thing that bumped it was `skill-author`
+Step 4, so every hand edit and every trim pass skipped the bump: on `docs/simplification-audit`, 17
+of the 40 skill directories that changed against `main` carried the same `metadata.version` as
+`main`, and `gitea` alone sat at six different values. Both validators checked presence and semver
+shape, never movement, so the field could not answer the question this ADR gives it — "did this
+change since I last read it". (Simplification audit finding 33.)
 
-**Decision.** `scripts/check-skill-version-bump.sh` runs as a pre-push hook on every push. For each
-skill directory under `plugins/*/.apm/skills/` that differs between the pushed ref and its
-merge-base with `main`, ignoring `tests/`, the pushed `metadata.version` must be strictly greater
-than `main`'s.
+`scripts/check-skill-version-bump.sh` now runs as a pre-push hook on every push, whatever the
+target branch. It takes its baseline from the merge-base of the pushed commit with `origin/main`
+(local `main` if `origin/main` does not resolve). For each skill directory under
+`plugins/*/.apm/skills/` that differs between the pushed commit and that merge-base, ignoring
+`tests/`, the pushed `metadata.version` must be strictly greater than the version the skill had at
+the merge-base — not the version on `main`'s current tip.
 
-- **Baseline is the merge-base with `main`, not the previous commit.** Readers only ever see
-  `main` — installs resolve against the default branch (ADR-0018) — so one bump per branch is what
-  the field owes them. A per-commit check would bump a skill once per commit and inflate the
-  number past meaning.
-- **"Greater", not "exactly one patch higher".** A second `skill-author` pass on the same branch
-  that bumps again still passes.
-- **`tests/` is excluded.** No agent loads it; a fixture-only change does not change the skill.
-- **Skills absent from either side are exempt.** New, renamed and merged skills start fresh under
-  the rules above; deleted skills have nothing to check.
-- **Every plugin is covered, `bin` included,** and the gate is repo-local — it is not exported
-  through `.pre-commit-hooks.yaml`.
+The baseline is the merge-base, not the previous commit. Readers only ever see `main` — installs
+resolve against the default branch (ADR-0018) — so one bump per branch is what the field owes
+them. A per-commit check would bump a skill once per commit and inflate the number past meaning.
+The rule is "greater", not "exactly one patch higher", so a second `skill-author` pass on the same
+branch that bumps again still passes. `tests/` is excluded because no agent loads it; a
+fixture-only change does not change the skill. Skills absent from either side are exempt: new,
+renamed and merged skills start fresh under the rules above, and deleted skills have nothing to
+check. Every plugin is covered, `bin` included, and the gate is repo-local — it is not exported
+through `.pre-commit-hooks.yaml`.
 
-**Considered options.** *Declare the field advisory* — cheapest, but concedes the field cannot do
-its job. *Drop the field* — rejected by this ADR already, and costlier now. *Check at commit
-time against `HEAD`* — rejected for the inflation above.
+The gate fails closed rather than passing when it has no trustworthy baseline: when neither
+`origin/main` nor `main` resolves, when the pushed commit shares no merge-base with it, and when
+only local `main` resolves and the pushed commit is that merge-base, since a local `main` the
+pushed commit already contains is no independent record of what shipped. It reads versions with
+`python3` and PyYAML and fails with a clear message if either is missing.
 
-**Consequences.** The 17 unbumped skills took a patch bump in the commit that added the gate.
-Like `check-release-needed`, the gate only fires on a local `git push` through pre-commit: a merge
-made with Gitea's merge button runs no local hooks and is not checked. A typo fix in a skill now
-costs a version bump; that is the rule working, not noise.
+Three alternatives were rejected. Declaring the field advisory is the cheapest, but concedes the
+field cannot do its job. Dropping the field was rejected by this ADR already, and costs more now.
+Checking at commit time against `HEAD` was rejected for the inflation described above.
+
+The 17 unbumped skills took a patch bump in the commit that added the gate. A typo fix in a skill
+now costs a version bump; that is the rule working, not noise.
+
+The gate differs from `check-release-needed` in when it runs: that hook acts only when pre-commit
+reports a push to `main`, so a manual `pre-commit run --hook-stage pre-push` skips it, while this
+gate runs there too and checks `HEAD`. The two hooks share both known gaps. A merge made with
+Gitea's merge button runs no local hooks, so it is not checked. And pre-commit's pre-push
+integration checks only the first ref with new commits in a multi-ref push (`git push origin a b`,
+`git push --all`): `_pre_push_ns` in pre-commit's `hook_impl.py` returns after that ref, so the
+other refs are pushed unchecked.
diff --git a/docs/spec/gates.md b/docs/spec/gates.md
index 3fbc1a0..ff5fbdb 100644
--- a/docs/spec/gates.md
+++ b/docs/spec/gates.md
@@ -36,6 +36,14 @@ blocked on cutting a premature tag — but it means `--hook-stage pre-push --all
 rehearsal of 8 hooks and a skip of the ninth. The script's own header records the same gap for
 a PR merged through Gitea's merge button, where no local push happens at all.
 
+A real push has a gap of its own. When one `git push` carries several refs
+(`git push origin a b`, `git push --all`), pre-commit runs the pre-push stage once, for the first
+ref that has new commits: `_pre_push_ns` in pre-commit's `hook_impl.py` returns as soon as it finds
+that ref. The two hooks that read the pushed ref — `check-release-needed` and
+`check-skill-version-bump` — therefore check only that ref, and the others are pushed unchecked.
+For `check-release-needed`, a multi-ref push whose first ref is not `main` never gates `main` at
+all. Push one ref at a time when the gate matters.
+
 ## The pre-push gate
 
 Nine hooks, grouped below by what they guard rather than by the order `.pre-commit-config.yaml` declares them in.
@@ -75,12 +83,17 @@ drift in generated text.
 |---|---|
 | `validate-marketplace` | `claude plugin validate --strict` on the root marketplace manifest |
 
+**Skill versioning** (every push, any branch)
+
+| Hook | Guards |
+|---|---|
+| `check-skill-version-bump` | fails if a skill directory changed since the pushed commit's merge-base with `main` without its `metadata.version` rising above the merge-base's (see [below](#check-skill-version-bump)) |
+
 **Release**
 
 | Hook | Guards |
 |---|---|
 | `check-release-needed` | on a real `git push` to `main` only — fails if files exposed via `.pre-commit-hooks.yaml` changed since the last tag. A no-op everywhere else, including under `pre-commit run --hook-stage pre-push` (see [the caveat above](#running-the-gates)) |
-| `check-skill-version-bump` | on every push — fails if a skill directory changed since the merge-base with `main` without its `metadata.version` rising (see [below](#check-skill-version-bump)) |
 
 Two of these shell out to `apm`: `apm-audit-ci` and `apm-pack-check-clean`. The second is a bare
 `apm …` entry and the first is a `bash -c` loop calling `apm` once per package, so without the CLI
@@ -89,20 +102,38 @@ the push dies with an unhelpful "command not found". Install with `apm-install`,
 
 ### `check-skill-version-bump`
 
-ADR-0022 makes `metadata.version` mandatory and a skill change carries a bump; `skill-size-check`
-only checks the field's presence and shape, so this hook holds the bump itself.
+ADR-0022 makes `metadata.version` mandatory and says a skill change carries a bump;
+`skill-size-check` only checks the field's presence and shape, so this hook holds the bump itself.
 
-- **Baseline is `git merge-base origin/main <pushed ref>`** (local `main` if `origin/main` does not
-  resolve). Readers install from `main`, so "changed" means changed against what `main` ships. The
-  remote branch tip is not the baseline: a second push would excuse an unbumped change the first
-  push already carried. With no `main` ref or no merge-base, the hook fails closed.
+- **It runs on every push and under a manual `pre-commit run --hook-stage pre-push`.** It does not
+  read `PRE_COMMIT_REMOTE_BRANCH`, so unlike `check-release-needed` the manual rehearsal really
+  checks it. The pushed commit is `PRE_COMMIT_TO_REF`, or `HEAD` when that is unset.
+- **Baseline is the merge-base of the pushed commit with `origin/main`** (local `main` if
+  `origin/main` does not resolve). The pushed version is compared with the skill's version *at
+  that merge-base*, not with `main`'s current tip. Readers install from `main`, so "changed" means
+  changed against the `main` the branch started from. The remote branch tip is not the baseline:
+  a second push would excuse an unbumped change the first push already carried.
+- **It fails closed when it has no trustworthy baseline:** neither `origin/main` nor `main`
+  resolves; there is no merge-base (shallow clone, unrelated history); or only local `main`
+  resolves and the pushed commit *is* the merge-base, so local `main` already contains the pushed
+  commit and says nothing independent about what shipped.
 - **A changed skill must end strictly above its baseline version**, compared numerically
   (`1.0.10` > `1.0.9`). Any bump size passes. A missing or non-`MAJOR.MINOR.PATCH` version at the
-  pushed ref fails. Skills absent at the baseline (new, renamed, merged) or at the pushed ref
-  (deleted) are exempt.
+  pushed ref fails, and so does a skill directory left without its `SKILL.md`. Versions are
+  ASCII-only with at most nine digits per part — stricter than `skill-size-check`'s shape check,
+  so a Unicode digit or an overflowing part cannot pass as a bump. Skills absent at the baseline
+  (new, renamed, merged) or at the pushed ref (deleted) are exempt.
+- **It also fails closed on read errors:** a pushed ref that does not resolve to a commit, or a
+  `SKILL.md` that `git show` or `python3` cannot read, stops the push with a read-failure message
+  rather than being reported as a missing version. Frontmatter that reads but does not parse
+  counts as an invalid version.
 - **`<skill>/tests/` is excluded**: no agent loads it, so a test-only change ships nothing.
-- **Known gap:** a PR merged through Gitea's merge button runs no local hook, the same gap
-  `check-release-needed` has.
+- **It needs `python3` and PyYAML** to read the frontmatter, and fails with a clear message if
+  either is missing, for the reasons in
+  [`python3` and PyYAML are hard requirements](#python3-and-pyyaml-are-hard-requirements).
+- **Known gaps, both shared with `check-release-needed`:** a PR merged through Gitea's merge button
+  runs no local hook; and a multi-ref push checks only its first ref with new commits (see
+  [Running the gates](#running-the-gates)).
 
 ## Skill and agent context gates (ADR-0020)
 
@@ -416,6 +447,8 @@ which is the exact vacuous-green failure the `python3` check exists to avoid. `p
 **Neither requirement generalises to every hook in this repo.** `check-rtk-prefix` needs `python3`
 but **not** PyYAML: it reads the markdown body and never touches frontmatter, so it has no scalar to
 fold.
+`check-skill-version-bump` needs both, for the same reason as `skill-size-check`: it parses
+`metadata.version` out of frontmatter.
 
 ## Agent files take the description gates, not the body gate