@event4u/agent-config 16.3.0 → 16.5.0

package/dist/agent-src/skills/command-routing/SKILL.md

@@ -99,11 +99,10 @@ Read `github.pr_reply_method` from `.agent-settings.yml`:
 | `create_review_comment` | `POST /repos/{owner}/{repo}/pulls/{number}/comments` with `{"body": "...", "in_reply_to": comment_id}` |
 | `auto` | Try `replies_endpoint` first. If it works → update setting. If 404/error → try `create_review_comment`. |
 
-### 2. Bot icon prefix
+### 2. No bot icon prefix
 
-Read `personal.pr_comment_bot_icon` from `.agent-settings.yml`:
-- `true` → prefix reply body with `🤖 `.
-- `false` or not set → no prefix.
+Never prefix the reply body with `🤖 ` — `🤖` is on the universal blacklist of
+[`no-decorative-emojis-in-git-surfaces`](../../rules/no-decorative-emojis-in-git-surfaces.md).
 
 ### 3. API call rules
 

package/dist/agent-src/skills/conventional-commits-writing/SKILL.md

@@ -46,10 +46,16 @@ imposing the shipped default there produces commits that read as foreign in
 | Tier | Source | Binding? |
 |---|---|---|
 | **1 — configured** | `commitlint.config.*` · `.gitmessage` · a `commit-msg` hook (husky / lefthook / `.git/hooks`) · `CONTRIBUTING.md` § Commits · a CI job that validates subjects · release automation that PARSES subjects (semantic-release, changesets, git-cliff, conventional-changelog) | yes — Class A, no approval needed |
+| **1b — declared** | `git.commit_format` in `.git-convention.yml`, the team's committed carrier at the repository root (ADR-283), where either value is a declaration — a team that commits `ticket-scope` chose it. Without the carrier, `git.commit_format: ticket-conventional` in a developer settings file, where only the non-default value declares anything: `ticket-scope` is also what a copied template carries, so it proves no choice. `agent-config git:convention show --key commit_format` names the file and the state; `malformed` or `invalid` declares nothing and is reported, never read as the default | yes — the user's own word, so no measurement and no ask |
 | **2 — measured + approved** | the consensus pass below, after the user says yes | yes, for this repository |
 | **3 — measured, unapproved** | the same pass before the user answers | **no — advisory**; report the mismatch, write Conventional |
 | **4 — default** | Conventional Commits | yes |
 
+`agent-config git:convention subject` validates against tiers 1b–4 only. A
+tier-1 commitlint config or `commit-msg` hook is named in a `note:` line as also
+running at commit, possibly stricter; the verb never infers what it accepts and
+never changes its exit for it.
+
 Release automation is the trap that makes tier 1 outrank tier 2 even when the
 history disagrees: a repo whose `git log` is 85 % `[JIRA-123] Fix thing` but
 whose manifest (`package.json`, `composer.json`, `pyproject.toml`, `Cargo.toml`)
@@ -60,6 +66,7 @@ Check for the parser before you trust the prevalence.
 ### 1. Look for tier 1 before measuring anything
 
 ```bash
+agent-config git:convention show --key commit_format   # tier 1b: valid from .git-convention.yml (either value), or ticket-conventional from a settings file
 ls commitlint.config.* .commitlintrc* .gitmessage .czrc 2>/dev/null
 git config --get commit.template
 ls .husky/commit-msg .git/hooks/commit-msg 2>/dev/null
@@ -88,144 +95,125 @@ The tier table above calls a confirmed hit Class A. That is exact for a machine
 read by a human judgement — treat a prose source as tier 1 in **precedence** and
 as a quoted sentence in the report, so the next reader can check it.
 
-### 2. Sample the history — native source per dimension
+### 2. Measure the history — `agent-config git:convention measure`
 
-Only **commit subjects** are reliably measurable from a clone. Branch names,
-PR titles and commit granularity have different native sources, and `unknown`
-is a valid, honest result for each:
+The measurement is code. `agent-config git:convention measure [--limit N]
+[--family F] [--json]` samples the history, classifies, caps, aggregates and
+judges it, and writes nothing. The sample size, the exclusions and every
+threshold it applies are the `measure` rows of the generated block in
+[`commit-subject`](../git-workflow/references/commit-subject.md) § The grammar,
+rendered from the module the verb runs — read them there, never from memory.
+What the verb reports, and why each choice is the one it makes:
 
-| Dimension | Native source | Measurable from a clone? |
+| Dimension | Native source | What `measure` does with it |
 |---|---|---|
-| Subject grammar | `git log` | yes |
-| Branch naming | forge refs / `git branch -r` | partially — merged branches are often deleted |
-| PR title shape | forge API (`gh pr list`) | only with forge access |
-| Granularity (atomic vs batched) | per-commit diff stat | **no** where the repo squash-merges — the evidence was destroyed at merge |
+| Subject grammar | `git log` on the trunk | classifies, caps, judges — the only dimension that can be **established** |
+| Branch naming | `git for-each-ref refs/remotes`, without `HEAD` and the default branch | proposes a `branch_pattern` among the shapes the reader accepts, or `no clear pattern` — merged branches are often deleted, so the sample is thin |
+| Update style | merge subjects that bring the default branch into a topic branch | **shown, never adopted** — `update_strategy` is a team decision, not a habit to copy |
+| PR title shape, granularity | forge API · per-commit diff stat | not measured — granularity is destroyed where the repository squash-merges |
 
-Resolve the trunk explicitly. `origin/HEAD` is set by `git clone` and is absent
-in a `git init` + `git remote add` checkout, in many CI checkouts and in
-worktrees, where the sample command would die rather than return nothing:
-
-```bash
-TRUNK=$(git symbolic-ref -q --short refs/remotes/origin/HEAD)
-[ -n "$TRUNK" ] || TRUNK=$(git rev-parse -q --verify origin/main >/dev/null && echo origin/main)
-[ -n "$TRUNK" ] || TRUNK=$(git rev-parse -q --verify origin/master >/dev/null && echo origin/master)
-[ -n "$TRUNK" ] || TRUNK=HEAD          # no remote at all: this checkout is the trunk
-git log "$TRUNK" --no-merges -n 200 --since='24 months ago' \
-  --pretty=format:'%H%x09%aN%x09%aE%x09%s'
-```
+**The trunk** is `origin/HEAD`, else `origin/main`, else `origin/master`, else
+`HEAD` — `origin/HEAD` is absent in a `git init` + `git remote add` checkout,
+in many CI checkouts and in worktrees, and a chain that silently falls through
+to an empty argument would hide that the resolution failed.
 
 **`--first-parent` is deliberately absent.** On a merge-commit workflow the
-trunk's first-parent chain is merge commits, which `--no-merges` then drops —
-the two together leave only what was committed straight to the trunk, which on
-a PR-based repo is nearly nothing. The convention lives in the feature commits,
-so walk the whole reachable history and exclude merges alone.
-
-The terminal `HEAD` matters as much as the first line: a chain that ends with an
-unset `TRUNK` hands `git log` an empty argument, which walks `HEAD` anyway —
-silently doing the right thing on a local repo while hiding that the resolution
-failed on one where it should not have.
-
-`%H` leads the format because two of the exclusions below need the commit, not
-just its subject. Drop, from the sample:
-
-- **bot authors** — `[bot]`, `dependabot`, `renovate`, `github-actions`,
-  `semantic-release`, `release-please`;
-- **automation subjects** — `^Revert `, `^Merge `, `chore(release)`,
-  `^Bump `, a bare `v?\d+\.\d+\.\d+`;
-- **bulk imports** — resolve with the hash the format carries:
-  `git show --stat --oneline <sha> | tail -1` and drop anything over 500 files.
-  Check this only for the handful of subjects that read like an import; running
-  it per commit costs 200 subprocesses to remove two rows.
-
-### 3. Classify each surviving subject
-
-Match in order, first hit wins. These are POSIX extended regular expressions,
-given outside a table because a markdown cell would need the alternation pipes
-escaped and `\|` in ERE is a literal pipe, not an alternation:
-
-```
-conventional      ^(build|chore|ci|docs|feat|fix|perf|refactor|revert|style|test)(\([^)]+\))?!?: 
-ticket-prefix     ^\[[A-Z][A-Z0-9]+-[0-9]+\][: ]|^[A-Z][A-Z0-9]+-[0-9]+[: ]
-gitmoji           ^:[a-z0-9_+-]+:[[:space:]]|^[^[:ascii:][:space:]]
-imperative-plain  ^[A-Z][a-z]+[[:space:]].*[^.]$
-other             everything else
-```
-
+trunk's first-parent chain is merge commits, which `--no-merges` then drops — the
+two together leave only what was committed straight to the trunk, which on a
+PR-based repo is nearly nothing. The convention lives in the feature commits.
+
+**Exclusions** — bot authors, automation subjects (reverts, merges, release
+chores, version bumps, bare versions) and bulk imports. The verb reads the file
+count of every sampled commit in the same `git log` call, so the bulk check costs
+no extra subprocess.
+
+**Classification** — the `family` rows of the same block, matched in order,
+first hit wins: `conventional`, `ticket-conventional`, `ticket-prefix`,
+`gitmoji`, `imperative-plain`, else `other`. `ticket-conventional` precedes
+`ticket-prefix`, which would otherwise swallow it as "ticket, then free text".
 `imperative-plain` is deliberately mechanical — capitalised first word, no
-trailing period — and does **not** test for the imperative mood. Mood needs a
+trailing period — and does **not** test for the imperative mood: mood needs a
 verb lexicon this procedure does not ship, and `Update` versus `Updated` is
-exactly the pair a lexicon-free test cannot separate. That is a stated limit of
-the classifier, not a gap to fill by guessing: a repo whose only distinction
-from Conventional is mood will read as `imperative-plain` either way, and the
-mood question belongs in the ask at step 6.
-
-`classifier_version` in step 7 names the revision of THIS section — the five
-patterns above plus the exclusions in step 2. It is a provenance stamp so a
-later measurement can be compared against a like one; it does not claim an
-executable classifier ships anywhere in the tree.
-
-Record the runner-up family too — a near-tie is itself the finding.
-
-### 4. Aggregate, capped per author, per half
-
-Raw commit counts let one prolific author or an unfiltered bot define the house
-style; one-vote-per-author lets a drive-by contributor with two commits veto it.
-Cap instead: split the sample into a newer and an older half by position first,
-then let **each author contribute at most 20 commits to each half**. Capping the
-sample as a whole would let one author's newest 20 fill their entire quota and
-empty the older half, which is precisely where step 5's coherence check has to
-look.
-
-Compute the dominant family's share over the **capped** total, and report that
-total — a percentage quoted against the raw eligible count describes a statistic
-nobody computed.
-
-With fewer than three authors the cap is **off**: there is no dominance to guard
-against, and a cap of 20 would put the single-maintainer bar's own `n ≥ 30`
-out of reach of its denominator.
-
-### 5. The evidence bar
-
-A measured convention is eligible to be proposed only when **all** hold:
-
-- **n ≥ 30** eligible commits after exclusions;
-- **≥ 80 %** capped-weighted share for the dominant family, over **≥ 3 human
-  authors** — or, with **one or two** authors, an uncapped **≥ 90 %** share over
-  n ≥ 30. Two authors take the stricter branch rather than falling between the
-  two: a two-person repo is a small-team repo, and leaving it unreachable by
-  both branches would make the measurement permanently advisory there;
-- **temporal coherence** — the same family is dominant in the newer half of
-  the sample *and* in the older half. If the halves disagree materially the
-  repository is **migrating**: abstain, name the change point, and use the
-  newer half's family only if it independently clears the bar.
-
-> The 80 % figure is the stricter of two council positions (2026-09-04,
-> anthropic + openai, 2/2 quorum: 70 % vs 80 % + a two-thirds-of-authors
-> clause). The per-author cap is why the separate author-share clause was not
-> also adopted — it already answers the dominance concern the clause existed
-> for. These are policy heuristics, not statistically derived thresholds.
-> **Revisit-if:** a labelled corpus of repositories shows a false-positive or
-> false-negative rate that a different bar would fix.
-
-### 6. Ask — never adopt silently
-
-Below the bar, or on a migrating repo: use Conventional Commits and say so
-**once**, in one line, when a commit is actually requested. Do not ask.
-
-At or above the bar, ask once, as numbered options (per `user-interaction`):
-name the family, the share **and the capped total it was computed over**, the
-author count, and — if a tier-1 parser exists — the tooling that would stop
-parsing.
-
-### 7. Persist the answer where the next session finds it
-
-Write the measurement as a Class-B convention card under
-`agents/memory/curated/conventions/quarantine/commit-subject.md`; the user's
-approval is what moves it to `approved/`. Carry `observed_n`,
-`dominant_share`, `author_count`, `sample_window`, `classifier_version`, and
-`confirm_against` — aggregates only, never per-author identities. Re-measure
-and re-review when the share falls below **70 %**, when a tier-1 source
-appears, or when the newer half's family changes.
+exactly the pair a lexicon-free test cannot separate. A repo whose only
+distinction from Conventional is mood reads as `imperative-plain` either way,
+and the mood question belongs in the ask at step 3. `classifier_version` in the
+verb's output and on the card names the revision of the families and exclusions
+(`ticket-conventional` was added 2026-10-06; re-measure an older card).
+
+**Aggregation, capped per author, per half.** Raw counts let one prolific author
+or an unfiltered bot define the house style; one vote per author lets a drive-by
+contributor with two commits veto it. The verb splits the sample into a newer and
+an older half by position first, then caps each author within each half —
+capping the sample as a whole would let one author's newest commits fill the
+quota and empty the older half, which is exactly where the coherence check has to
+look. Shares are over the **capped** total, and the verb reports that total; a
+percentage quoted against the raw eligible count describes a statistic nobody
+computed. With fewer than three authors the cap is off and the stricter uncapped
+bar applies: there is no dominance to guard against, and a cap would put the
+minimum sample out of reach of its own denominator. Two authors take the stricter
+branch rather than falling between the two.
+
+**Temporal coherence.** The same family must lead the newer **and** the older
+half. Halves that disagree mean the repository is **migrating**: the verb names
+both families and establishes the newer half's family only if the newer half
+clears the bar on its own (`migrating` in the output).
+
+**Verdict.** At or above the bar the verb names the established family; below it,
+the reasons and the **two strongest** families with a grammar. Record the
+runner-up either way — a near-tie is itself the finding.
+
+> The share bar is the stricter of two council positions (2026-09-04,
+> anthropic + openai, 2/2 quorum); the per-author cap is why a separate
+> author-share clause was not also adopted — it already answers the dominance
+> concern that clause existed for. These are policy heuristics, not statistically
+> derived thresholds. **Revisit-if:** a labelled corpus of repositories shows a
+> false-positive or false-negative rate that a different bar would fix.
+
+### 3. Ask once — never adopt silently
+
+Only an **interactive** `/commit` asks, and only while `git:convention show --key commit_format`
+prints `no convention established` — no declaration and no approved card. It
+asks once, as numbered options (per `user-interaction`), naming the family, the
+share **and the capped total it was computed over**, the author count, and — if a
+tier-1 parser exists — the tooling that would stop parsing:
+
+- **at or above the bar** — the established family (recommended) vs Conventional
+  Commits;
+- **below the bar** — the two strongest families vs Conventional Commits, with
+  the reasons the verb gave.
+
+`/commit:in-chunks` never asks: it reports the measurement in its summary and
+uses the card, or the default. Until the user answers, a measurement is tier 3 —
+advisory.
+
+### 4. Persist the answer where the next session finds it
+
+The answer is the approval, so `/commit` writes the card straight to
+`agents/memory/curated/conventions/approved/commit-subject.md` — **also when the
+answer is Conventional Commits**, so the question is never asked again.
+`git:convention measure --family <answer>` prints the card's content: a Class-B
+convention card carrying `dominant_family`, `observed_n`, `dominant_share`,
+`author_count`, `sample_window`, `classifier_version`, `confirm_against` and
+`ticket_keys` — aggregates only, never per-author identities. A measurement the
+agent records without an answer goes to `quarantine/` instead. `ticket_keys`
+lists the approved project keys (`ticket_keys: [DEV, OPS]`): it is filled from the
+user's answer when `/commit` meets a key not on it, and a commitlint config's
+`issuePrefixes` are offered as the proposal, never written silently. Re-measure
+and re-review when the share falls below **70 %**, when a tier-1 source appears,
+or when the newer half's family changes.
+
+### 5. The team file is a human's commit
+
+A card is one developer's memory; `.git-convention.yml` is the team's
+declaration and outranks it (tier 1b). When the chosen family maps to a
+`git.commit_format` value (`conventional` → `ticket-scope`, `ticket-conventional`
+→ `ticket-conventional`) or a `branch_pattern` was proposed, `measure` — and
+`/commit` after the answer — prints the file's ready-to-commit content.
+`update_strategy` is never in it. The file is class C: the agent prints it and a
+human creates and commits it. The one exception is creation on the user's
+explicit yes this turn, through `agent-config git:convention init --yes` — once,
+never over an existing carrier, never `update_strategy`, never committed by the
+verb; the agent never writes or edits the file by hand.
 
 ## What a measurement may never lower
 
@@ -274,18 +262,15 @@ If yes:
 - Suggest splitting into multiple commits
 - Or choose the dominant net effect for squash merge title
 
-### 3. Choose scope
+### 3. Place the ticket, then choose the scope
 
-Add a scope only if it improves clarity:
-
-- Jira ticket ID: `DEV-1234`
-- Module/area: `api`, `auth`, `skills`, `rules`, `ci`
+Read the ticket with `agent-config git:convention ticket` and place it per [`commit-subject`](../git-workflow/references/commit-subject.md) § Placing the ticket — under `ticket-conventional` **never in the scope**.
 
 ### 4. Write the description
 
 - State the intent clearly
 - Avoid generic filler (`update stuff`, `fix things`)
-- Stay concise — max 72 chars total for first line
+- Stay concise — max 72 chars total for first line (the ticket prefix counts)
 - Imperative mood: "add", "fix", "remove" — not "added", "fixed", "removed"
 
 ### 5. Check for breaking change
@@ -316,16 +301,18 @@ Or add `BREAKING CHANGE:` in the commit body/footer.
 
 ## Procedure: Generate squash merge title
 
+Only for a squash merge — rebase-and-merge keeps every commit (`/pr:merge` § 9 reads the method from the forge).
+
 1. Read all commits in the PR
 2. Identify the **net effect** — what does the PR accomplish overall?
-3. Write a single Conventional Commit message summarizing the net effect
+3. Write a single subject in the convention in force (ticket placed per § Place the ticket) summarizing the net effect
 4. Do not list every internal commit — summarize
 
 ## Output format
 
 1. The convention in force and the tier that established it — `configured
-   (commitlint.config.js)`, `approved (ticket-prefix, 84% of 137)`, or
-   `default (Conventional Commits)`
+   (commitlint.config.js)`, `declared (git.commit_format: ticket-conventional)`,
+   `approved (family ticket-prefix, 84% of 137)`, or `default (Conventional Commits)`
 2. Recommended commit message(s)
 3. Brief rationale for type choice
 4. Split suggestion if the change should be multiple commits
@@ -372,6 +359,7 @@ to every commit message you author.
   advisory, and silence is not approval
 - Do NOT let prevalence lift a `never` or `explicit-only` floor
 - Do NOT use vague messages: `update stuff`, `fix bug`, `changes`
+- Do NOT put a ticket id in the scope when `git.commit_format` is `ticket-conventional`
 - Do NOT use `refactor` for bug fixes
 - Do NOT use `chore` for meaningful behavior changes
 - Do NOT hide multiple unrelated concerns in one message

package/dist/agent-src/skills/corpus-grounding/ATTRIBUTION.md

@@ -0,0 +1,39 @@
+# Attribution — the ported grounding engine
+
+This file exists because of where the code ships, not because the obligation
+is new. `src/skills/design-intelligence/ATTRIBUTION.md` already covered both
+the vendored corpus **and** the ported engine. The corpus travels with the
+`frontend-design` pack; the engine sources below travel with **this** skill, in
+`engineering-base`. An install that takes `engineering-base` without
+`frontend-design` therefore carried MIT-derived code and no notice, which is
+the one thing the MIT obligation forbids. The notice is reproduced here so it
+travels with the files it covers.
+
+## Covered files
+
+- `scripts/bm25_search.ts`
+- `scripts/decision_engine.ts`
+- `scripts/ground.ts`
+- `scripts/schema_validator.ts`
+
+Ported from `nextlevelbuilder/ui-ux-pro-max-skill` @
+`b7e3af80f6e331f6fb456667b82b12cade7c9d35` (Python sources `core.py`,
+`search.py`, `design_system.py`), de-duplicated and with slide-only paths
+stripped.
+
+## MIT
+
+Copyright (c) 2024 **Next Level Builder**.
+
+The Python engine sources (`core.py`, `search.py`, `design_system.py`) this
+skill's scripts are ported from are MIT-licensed. Obligation: **retain this
+notice**.
+
+## Not covered here
+
+The tabular design-knowledge corpus (`ui-reasoning`, `products`, `colors`,
+`styles`, `typography`, `charts`, `landing`, `icons`, `ux-guidelines`,
+`react-performance`, `app-interface`, `motion`, `stacks/*`), the Apache-2.0
+`ui-styling`-derived material, and the running modifications log stay in
+`design-intelligence/ATTRIBUTION.md` — they ship with that pack, and this file
+does not duplicate them.

package/dist/agent-src/skills/corpus-grounding/SKILL.md

@@ -31,8 +31,14 @@ execution:
 Engine provenance: ported from `nextlevelbuilder/ui-ux-pro-max-skill`
 @ `b7e3af80f6e331f6fb456667b82b12cade7c9d35` (MIT, last checked
 2026-06-07) — BM25 de-duplicated, slide-only paths stripped, every
-frontend-hardcoded axis moved into the manifest. Full license obligations:
-[`design-intelligence/ATTRIBUTION.md`](../design-intelligence/ATTRIBUTION.md).
+frontend-hardcoded axis moved into the manifest. The obligations split by what
+ships where: the MIT notice for the **engine sources under `scripts/`** travels
+with this skill, in [`ATTRIBUTION.md`](ATTRIBUTION.md) beside it, so an install
+that takes `engineering-base` without `frontend-design` still carries the
+notice its code is under. The vendored **corpus** and the Apache-2.0
+`ui-styling` material travel with `frontend-design`, in
+[`design-intelligence/ATTRIBUTION.md`](../design-intelligence/ATTRIBUTION.md),
+which remains the full record.
 
 ## When to use
 

package/dist/agent-src/skills/database/data/manifest.json

@@ -29,7 +29,7 @@
   "refresh_cadence": "quarterly",
   "upstream": {
     "repo": "derived: PostgreSQL 18 + MySQL 9.7 reference docs (previously PostgreSQL 16 + MySQL 8)",
-    "sha": "re-derived 2026-09-27 against PostgreSQL 18.x / MySQL 9.7: all 12 rows of query-tuning.csv still apply as written on both engines. Three rows gained a version-scoped annotation for a genuinely new capability rather than a rewrite \u2014 row 1 (Slow WHERE on large table) notes PG18's B-tree skip scan, which relaxes the equality-first composite-index rule on PG18+ only; row 5 (JOIN explosion) notes MySQL 9.7's Hypergraph Optimizer, now GA in Community Edition, improving join-order choices automatically; row 12 (JSON column filtering) notes that PG18 flips the GENERATED-column default from STORED to VIRTUAL, and VIRTUAL columns are not indexable, so STORED must now be explicit. PostgreSQL 16.x and MySQL 8.4 remain supported and the patterns apply there unchanged",
-    "last_checked": "2026-09-18"
+    "sha": "PostgreSQL 18.x / MySQL 9.7 reference documentation (PostgreSQL 16.x and MySQL 8.4 also supported)",
+    "last_checked": "2026-10-10"
   }
 }

package/dist/agent-src/skills/database/data/query-tuning.csv

@@ -1,13 +1,13 @@
-Symptom,Keywords,Root Cause,Strategy,Code Good,Anti-Pattern,Verification,Severity
-Slow WHERE on large table,slow query where filter full scan seq scan,Sequential scan despite selective predicate,B-tree index on the filtered column(s); composite index ordered by equality-first then range (PostgreSQL 18+: skip scan lets the planner use later columns without a leading-column equality predicate — MySQL and pre-18 Postgres still need equality-first ordering),"CREATE INDEX idx_orders_status_created ON orders (status, created_at)",Indexing low-cardinality boolean alone; composite ordered range-first,EXPLAIN (ANALYZE) shows Index Scan (PG 18+: or Index Searches for a skip scan); rows estimate ≈ actual,HIGH
-ORDER BY ... LIMIT slow,order by limit sort top n latest,Sort node over the whole table for a top-N query,Index matching the ORDER BY direction (including DESC/NULLS) so LIMIT walks the index,CREATE INDEX idx_events_created_desc ON events (created_at DESC),OFFSET pagination for infinite feeds (use keyset),Plan shows Index Scan Backward / no Sort node,HIGH
-N+1 query loop,n+1 loop per row lazy load orm repeated query,Hundreds of identical queries differing only by id,Eager-load relations / batch with IN(...) or JOIN; dataloader pattern at the API layer,SELECT * FROM comments WHERE post_id IN (...),Fixing with a cache instead of the access pattern,Query count per request flat vs item count,HIGH
-LIKE '%term%' search,like contains search wildcard fulltext,Leading-wildcard LIKE can't use B-tree,Trigram index (pg_trgm GIN) or full-text search (tsvector + GIN); dedicated engine beyond ~10M rows,CREATE INDEX idx_users_name_trgm ON users USING gin (name gin_trgm_ops),Forcing LOWER() without an expression index,Plan shows Bitmap Index Scan on the GIN index,MEDIUM
-JOIN explosion,join many tables slow hash join nested loop,Mis-estimated join order / missing join keys index,Index every FK used in joins; verify row estimates; break monster joins into CTE/temp stages when estimates collapse (MySQL 9.7: the Hypergraph Optimizer is now GA in Community Edition and can materially improve join-order choices on its own — verify estimates before assuming a manual rewrite is still needed),ANALYZE tables; index fk columns,Disabling nested loops globally; hint-chasing before stats are fresh,Estimates within ~10x of actuals at every join node,HIGH
-Aggregation over millions of rows,group by count sum report aggregate slow,Repeated full-table aggregation for dashboards,Pre-aggregate: materialized view / rollup table refreshed on schedule or trigger; partial indexes for hot subsets,CREATE MATERIALIZED VIEW daily_rev AS ... REFRESH CONCURRENTLY,Caching in app memory only (cold-start storms),"Dashboard reads hit the rollup, not base tables",MEDIUM
-Pagination deep offset,offset page 10000 slow deep pagination,OFFSET N scans+discards N rows,Keyset/cursor pagination on a unique ordered key,"WHERE (created_at,id) < (:c,:i) ORDER BY created_at DESC,id DESC LIMIT 20",Hiding offset behind a 'page' param and capping at 1000 as a fix,p99 latency flat across page depth,HIGH
-Write contention / lock waits,deadlock lock wait contention update hot row,"Hot-row updates serialize (counters, balances)",Row-level design: ledger inserts over UPDATE-in-place; batch with SKIP LOCKED queues; keep transactions short,SELECT ... FOR UPDATE SKIP LOCKED for workers,Retrying deadlocks blindly without ordering writes,Lock wait time ~0 under load test; deadlock rate 0,HIGH
-Missing/stale statistics,wrong plan estimate stale statistics analyze,Optimizer estimates wildly off → bad plans,Auto-analyze tuned for big tables; manual ANALYZE after bulk loads; extended statistics for correlated columns,"CREATE STATISTICS s_dist (dependencies) ON city,country FROM addr",Hinting every query instead of fixing stats,Plan estimates ≈ actuals after ANALYZE,MEDIUM
-Over-indexing writes,insert slow many indexes write amplification,Each extra index taxes every INSERT/UPDATE,Audit with index-usage stats; drop unused; covering index to replace several near-duplicates,pg_stat_user_indexes idx_scan = 0 candidates,Adding an index per slow query without retiring old ones,Write p95 recovers; no unused-index rows,MEDIUM
-SELECT * over wide rows,select star wide row toast network payload,Fetching unused columns (TOAST/LOB) dominates IO,Project needed columns; covering index enables index-only scans,"SELECT id,status,total FROM orders ...",Compressing payloads instead of trimming the select list,Plan shows Index Only Scan where applicable,LOW
-JSON column filtering,json jsonb attribute filter schemaless,Filtering inside JSON without index = full scan,"GIN index on jsonb path ops, or generated column + B-tree for hot keys",ALTER TABLE t ADD COLUMN status text GENERATED ALWAYS AS (doc->>'status') STORED; CREATE INDEX ...; -- PG18+: STORED must be explicit (default flipped to VIRTUAL — not indexable),Stuffing relational data into one JSON blob; relying on the pre-PG18 STORED default,Plan uses the GIN/generated-column index,MEDIUM
+Symptom,Keywords,Root Cause,Strategy,Code Good,Anti-Pattern,Verification,Severity
+Slow WHERE on large table,slow query where filter full scan seq scan,Sequential scan despite selective predicate,B-tree index on the filtered column(s); composite index ordered by equality-first then range (PostgreSQL 18+: skip scan lets the planner use later columns without a leading-column equality predicate — MySQL and pre-18 Postgres still need equality-first ordering),"CREATE INDEX idx_orders_status_created ON orders (status, created_at)",Indexing low-cardinality boolean alone; composite ordered range-first,EXPLAIN (ANALYZE) shows Index Scan (PG 18+: or Index Searches for a skip scan); rows estimate ≈ actual,HIGH
+ORDER BY ... LIMIT slow,order by limit sort top n latest,Sort node over the whole table for a top-N query,Index matching the ORDER BY direction (including DESC/NULLS) so LIMIT walks the index,CREATE INDEX idx_events_created_desc ON events (created_at DESC),OFFSET pagination for infinite feeds (use keyset),Plan shows Index Scan Backward / no Sort node,HIGH
+N+1 query loop,n+1 loop per row lazy load orm repeated query,Hundreds of identical queries differing only by id,Eager-load relations / batch with IN(...) or JOIN; dataloader pattern at the API layer,SELECT * FROM comments WHERE post_id IN (...),Fixing with a cache instead of the access pattern,Query count per request flat vs item count,HIGH
+LIKE '%term%' search,like contains search wildcard fulltext,Leading-wildcard LIKE can't use B-tree,Trigram index (pg_trgm GIN) or full-text search (tsvector + GIN); dedicated engine beyond ~10M rows,CREATE INDEX idx_users_name_trgm ON users USING gin (name gin_trgm_ops),Forcing LOWER() without an expression index,Plan shows Bitmap Index Scan on the GIN index,MEDIUM
+JOIN explosion,join many tables slow hash join nested loop,Mis-estimated join order / missing join keys index,Index every FK used in joins; verify row estimates; break monster joins into CTE/temp stages when estimates collapse (MySQL 9.7: the Hypergraph Optimizer ships in Community Edition but is OFF by default — optimizer_switch reads hypergraph_optimizer=off; switch it on per session or statement to let it choose join order, then verify estimates before assuming a manual rewrite is still needed),ANALYZE tables; index fk columns,Disabling nested loops globally; hint-chasing before stats are fresh,Estimates within ~10x of actuals at every join node,HIGH
+Aggregation over millions of rows,group by count sum report aggregate slow,Repeated full-table aggregation for dashboards,Pre-aggregate: materialized view / rollup table refreshed on schedule or trigger; partial indexes for hot subsets,CREATE MATERIALIZED VIEW daily_rev AS ... REFRESH CONCURRENTLY,Caching in app memory only (cold-start storms),"Dashboard reads hit the rollup, not base tables",MEDIUM
+Pagination deep offset,offset page 10000 slow deep pagination,OFFSET N scans+discards N rows,Keyset/cursor pagination on a unique ordered key,"WHERE (created_at,id) < (:c,:i) ORDER BY created_at DESC,id DESC LIMIT 20",Hiding offset behind a 'page' param and capping at 1000 as a fix,p99 latency flat across page depth,HIGH
+Write contention / lock waits,deadlock lock wait contention update hot row,"Hot-row updates serialize (counters, balances)",Row-level design: ledger inserts over UPDATE-in-place; batch with SKIP LOCKED queues; keep transactions short,SELECT ... FOR UPDATE SKIP LOCKED for workers,Retrying deadlocks blindly without ordering writes,Lock wait time ~0 under load test; deadlock rate 0,HIGH
+Missing/stale statistics,wrong plan estimate stale statistics analyze,Optimizer estimates wildly off → bad plans,Auto-analyze tuned for big tables; manual ANALYZE after bulk loads; extended statistics for correlated columns,"CREATE STATISTICS s_dist (dependencies) ON city,country FROM addr",Hinting every query instead of fixing stats,Plan estimates ≈ actuals after ANALYZE,MEDIUM
+Over-indexing writes,insert slow many indexes write amplification,Each extra index taxes every INSERT/UPDATE,Audit with index-usage stats; drop unused; covering index to replace several near-duplicates,pg_stat_user_indexes idx_scan = 0 candidates,Adding an index per slow query without retiring old ones,Write p95 recovers; no unused-index rows,MEDIUM
+SELECT * over wide rows,select star wide row toast network payload,Fetching unused columns (TOAST/LOB) dominates IO,Project needed columns; covering index enables index-only scans,"SELECT id,status,total FROM orders ...",Compressing payloads instead of trimming the select list,Plan shows Index Only Scan where applicable,LOW
+JSON column filtering,json jsonb attribute filter schemaless,Filtering inside JSON without index = full scan,"GIN index on jsonb path ops, or generated column + B-tree for hot keys",ALTER TABLE t ADD COLUMN status text GENERATED ALWAYS AS (doc->>'status') STORED; CREATE INDEX ...; -- PG18+: STORED must be explicit (default flipped to VIRTUAL — not indexable),Stuffing relational data into one JSON blob; relying on the pre-PG18 STORED default,Plan uses the GIN/generated-column index,MEDIUM

package/dist/agent-src/skills/design-intelligence/references/integration-mapping.md

@@ -62,16 +62,19 @@ shape; 3–7 → one diagram; 8+ → split or summarize, never one dense picture
 | a **cycle** / loop / lifecycle | — | a **stepper widget**, never a hand-drawn ring |
 | a **DB schema / ERD** / entity relations | — | **mermaid**, never hand-placed SVG |
 
-### Geometric pre-checks (run BEFORE finalizing an SVG/diagram)
-
-Ranked by failure rate — procedures, not constants:
-
-1. **viewBox safety** — compute the lowest + rightmost element (plus a buffer)
-   and set the viewBox from that; never assume the default fits.
-2. **arrow-through-box trace** — trace every arrow's path and confirm it does
-   not cross through an unrelated box before drawing it.
-3. **box-width-from-longest-label** — size each box from its longest label
-   before placing it, so text never overflows.
+### Drawing the diagram itself
+
+The verb routing above and the noun-count triage stay here — they decide WHAT to
+draw. **How** to draw it, and the check that the drawing is right, belong to
+[`svg-authoring`](../../svg-authoring/SKILL.md).
+
+The three checks this section used to carry — viewBox safety, the
+arrow-through-box trace, and sizing a box from its longest label — were author
+self-checks with no witness, and a published example in this very file violated
+the viewBox rule it stated. They are now readings the asset profile takes in a
+real engine: every painted element against the viewBox, and every `<text>`
+against its declared box. A self-check that nothing measures is the failure
+mode, not the three rules.
 
 (Reference-only: any color/easing/frame values come from the consumer's tokens
 or a maintained upstream — this skill vendors no drawn-asset corpus.)

package/dist/agent-src/skills/design-system-capture/SKILL.md

@@ -30,9 +30,6 @@ execution:
 > **This skill captures *usage decisions*:** "we use 8px radius for
 > elevated surfaces", "all destructive actions require typed confirmation".
 > They are complementary, not redundant.
->
-> Gated by `tokens.rich_skills` in `.agent-settings.yml` (default `on`).
-> If `off`, skip rich context loading and work from brief alone.
 
 ## Why this skill is rich
 

package/dist/agent-src/skills/estimate-ticket/SKILL.md

@@ -80,7 +80,7 @@ If the ticket scores low on estimability (vague AC, no out-of-scope,
 no mention of dependencies), set `uncertainty = Underspecified` and
 emit a single-line redirect to `/refine-ticket` at the top of the output.
 
-A **cross-source discrepancy** (per [`cross-source-consistency`](../../rules/cross-source-consistency.md), gated by `consistency.cross_source`) — ticket text contradicting an attached mockup, or a clearly-needed-but-unstated behavior (weekend/holiday shift) — is not estimable as-is: it hides an open scope decision. Raise `uncertainty` and name the discrepancy in the redirect rather than sizing over a contradiction.
+A **cross-source discrepancy** (per [`cross-source-consistency`](../../rules/cross-source-consistency.md)) — ticket text contradicting an attached mockup, or a clearly-needed-but-unstated behavior (weekend/holiday shift) — is not estimable as-is: it hides an open scope decision. Raise `uncertainty` and name the discrepancy in the redirect rather than sizing over a contradiction.
 
 ### 3. Apply sizing heuristics
 

package/dist/agent-src/skills/fe-design/references/design-patterns.md

@@ -235,12 +235,59 @@ silently does nothing and a shared-element route change gets hand-rolled.
 - **`transform-origin`** decides where a `scale()` grows from. A menu or popover
   scaling in should originate at its trigger, not at its own centre, or it reads
   as arriving from nowhere. Scale in from ~0.96, never from `scale(0)`: growing
-  from nothing is an entrance, and steps 1-2 asked for feedback.
+  from nothing is an entrance, and steps 1-2 asked for feedback. This line is
+  about an HTML box. On an SVG child the same declaration resolves against the
+  viewBox instead, which is the first bullet of 6b below.
 - **`@media (hover: hover)`** gates any hover-carried motion. Without it a touch
   device fires the hover treatment on tap and leaves it stuck, because there is
   no pointer to leave.
 
-All six still obey step 3's bands, step 2's single `ease-out` curve, and the
+**6b. Which mechanism carries it inside an SVG?**
+
+An SVG child is not an HTML box, and the asset is often delivered somewhere the
+page's own CSS cannot reach. Four mechanisms differ there, and the delivery mode
+decides which of them is even available. Mechanism only: none of these states a
+duration, a band or a curve, and an SVG animation is banded by steps 1-3 like
+any other.
+
+- **`transform-box: fill-box`** is required before any CSS `transform` or
+  `transform-origin` on an SVG child means what it reads as. The initial
+  reference box is the nearest viewport — the `viewBox` — not the element's own
+  bounds, so `transform-origin: center` on a shape away from the origin scales
+  it about the canvas center and translates it as a side effect. The shape still
+  moves, nothing errors, and the value looks correct in the stylesheet, which is
+  why this is a silent failure rather than a visible one.
+- **`pathLength="1"`** normalises a path's length for a draw-on, so
+  `stroke-dasharray` and `stroke-dashoffset` are written as fractions of the
+  stroke instead of as units of a geometry nobody measured. A dash guessed too
+  long leaves the stroke invisible for the opening part of its run and then
+  snaps; guessed too short, it draws and stops early. Set the attribute and the
+  guess disappears.
+- **`keySplines`** takes control-point values inside `[0,1]`. An overshoot is
+  not expressible there and an out-of-range value is not clamped — the element
+  holds its start value for the whole run instead of easing, while an in-range
+  twin finishes normally. Step 2's bounce prohibition already rules the intent
+  out; this is the mechanism refusing it as well.
+- **A path morph** interpolates only between two `d` values carrying the same
+  number of commands, of the same type, in the same order. Anything else is a
+  discrete swap wearing an animation's syntax. Author both paths against one
+  command list rather than simplifying each in isolation. (The CSS `d` property
+  is deliberately not addressed here — its cross-engine behavior is unmeasured
+  in this tree.)
+- **Delivery mode picks the mechanism, and picks the reduced-motion route with
+  it.** A self-contained asset animates with internal CSS keyframes inside
+  `@media (prefers-reduced-motion: no-preference)`, which satisfies step 3's
+  requirement from inside the file. Reach for SMIL (`<animate>`,
+  `<animateTransform>`) only for what CSS cannot carry — and then the
+  reduced-motion route cannot be a CSS rule, because a page-level
+  `svg * { animation: none !important }` leaves an `<animate>` running. Use a
+  static twin selected by the media query, or `pauseAnimations()` under
+  `matchMedia`. An animated SVG delivered through `<img>` or a CSS background
+  cannot be reached by the host's CSS at all, so the host supplies the static
+  route — `<picture><source media="(prefers-reduced-motion: reduce)">` — or the
+  asset carries its own internal media query.
+
+All of them still obey step 3's bands, step 2's single `ease-out` curve, and the
 `@media (prefers-reduced-motion: reduce)` requirement above — a newer mechanism
 is not an exemption from the authority, it is another way of spending it.
 

package/dist/agent-src/skills/feature-planning/SKILL.md

@@ -28,7 +28,7 @@ Do NOT use when:
 
 ## Procedure: Plan a feature
 
-1. **Gather requirements** — What problem does this solve? Who benefits? What's the expected outcome? Run a **cross-source discrepancy scan** (per [`cross-source-consistency`](../../rules/cross-source-consistency.md), gated by `consistency.cross_source`): where the source ticket carries text + an attachment/mockup, or leaves a clearly-needed behavior unstated (weekend/holiday shift, empty/error state, timezone), surface the conflict/gap as an *Open question* before drafting phases — an inferred behavior is a scope expansion to confirm, not to bake in. See [`cross-source-consistency-mechanics`](../../guidelines/agent-infra/cross-source-consistency-mechanics.md).
+1. **Gather requirements** — What problem does this solve? Who benefits? What's the expected outcome? Run a **cross-source discrepancy scan** (per [`cross-source-consistency`](../../rules/cross-source-consistency.md)): where the source ticket carries text + an attachment/mockup, or leaves a clearly-needed behavior unstated (weekend/holiday shift, empty/error state, timezone), surface the conflict/gap as an *Open question* before drafting phases — an inferred behavior is a scope expansion to confirm, not to bake in. See [`cross-source-consistency-mechanics`](../../guidelines/agent-infra/cross-source-consistency-mechanics.md).
 2. **Analyze scope** — Which modules, routes, models, services are affected?
 3. **Write the plan** — Create a feature plan document (see template below).
 4. **Verify** — Confirm all affected areas are identified, no missing dependencies.