@event4u/agent-config 7.5.0 → 8.0.0

package/dist/agent-src/skills/complexity-first-planning/SKILL.md

@@ -1,7 +1,6 @@
 ---
 name: complexity-first-planning
 description: "Use when staging multi-component or uncertain work — tackle the load-bearing unknown first (risk-first decomposition), not the easy parts first."
-source: package
 domain: engineering
 status: active
 model_tier: medium
@@ -43,10 +42,12 @@ unknown to resolve), or when the user has already fixed the sequence.
 
 ## Procedure
 
-1. **Inspect and name the unknowns.** Read the affected components first, then
-   list which carry real uncertainty (technical feasibility, an unverified
-   integration, an ambiguous requirement) — analyze the existing system before
-   planning any change.
+1. **Inspect and name the unknowns.** Read the affected components first
+   (start from the **repo** slot of the
+   [context-spine](../../../docs/contracts/context-spine.md) when present),
+   then list which carry real uncertainty (technical feasibility, an
+   unverified integration, an ambiguous requirement) — analyze the existing
+   system before planning any change.
 2. **Assess and rank by load-bearing risk.** The load-bearing unknown is the one
    whose failure invalidates the most dependent work — not the one that is merely hard.
 3. **Resolve it first.** Spike / probe / prototype the load-bearing unknown

package/dist/agent-src/skills/condense-memory/SKILL.md

@@ -36,13 +36,13 @@ Use when:
 ## Procedure
 
 1. **Analyse target first.** Before any write, **inspect** target with `view` or `wc -l` to confirm it is always-loaded memory file (`AGENTS.md`, `CLAUDE.md`, `.cursorrules`, `GEMINI.md`, `.windsurfrules`), not generated, and has prose paragraphs to condense (pointer-only Thin-Root file may net near-zero). Skip rest of procedure if any check fails.
-2. **Check denylist gate.** Run `python3 scripts/condense_memory.py <path> --check` — exit 0 = safe; exit 2 = denylist hit, stop and surface refusal.
+2. **Check denylist gate.** Run `./scripts-run src/scripts/condense_memory <path> --check` — exit 0 = safe; exit 2 = denylist hit, stop and surface refusal.
 3. **Record baseline.** `wc -c <path>` — capture pre-condensation char count for commit message.
-4. **Condense.** `python3 scripts/condense_memory.py <path>`. Script writes `<path>.original.md` (verbatim backup) and rewrites `<path>` with `original_sha256:` + `condensed_at:` frontmatter.
+4. **Condense.** `./scripts-run src/scripts/condense_memory <path>`. Script writes `<path>.original.md` (verbatim backup) and rewrites `<path>` with `original_sha256:` + `condensed_at:` frontmatter.
 5. **Inspect diff.** Eyeball every Iron-Law fence, numbered-options block, code fence, backtick span, `❌`/`⚠️`/`✅` line, and frontmatter pair — all must be byte-identical. Body prose may have lost articles (`the`/`a`/`an`) and auxiliaries (`is`/`are`/`was`/`be`/`that`/`which`).
-6. **Validate idempotency.** Re-run `python3 scripts/condense_memory.py <path>` — clean re-run is no-op (body hash matches). Non-zero exit = stop, escalate.
+6. **Validate idempotency.** Re-run `./scripts-run src/scripts/condense_memory <path>` — clean re-run is no-op (body hash matches). Non-zero exit = stop, escalate.
 7. **Commit both files together.** `<path>` and `<path>.original.md` ship as pair. Backup is rollback path; never commit one without other.
-8. **Rollback path.** If readability fails review at step 5: `python3 scripts/condense_memory.py <path> --decondense` restores backup and deletes `.original.md`.
+8. **Rollback path.** If readability fails review at step 5: `./scripts-run src/scripts/condense_memory <path> --decondense` restores backup and deletes `.original.md`.
 
 ## Output format
 
@@ -57,7 +57,7 @@ Do **not** narrate algorithm, grammar rules, or carve-out theory — rule and th
 
 ## Carve-outs — byte-for-byte preserved
 
-Mirrors seven carve-out classes in [`telegraph-speak`](../../rules/telegraph-speak.md). Condensation engine in [`scripts/condense_memory.py`](../../../scripts/condense_memory.py) preserves:
+Mirrors seven carve-out classes in [`telegraph-speak`](../../rules/telegraph-speak.md). Condensation engine in [`src/scripts/condense_memory.ts`](../../../src/scripts/condense_memory.ts) preserves:
 
 1. **Triple-backtick fences** — any language, any depth.
 2. **Numbered-options lines** — `^>?\s*\d+\.\s` plus `**Recommendation:**` / `**Empfehlung:**` label.
@@ -67,7 +67,7 @@ Mirrors seven carve-out classes in [`telegraph-speak`](../../rules/telegraph-spe
 6. **Frontmatter blocks** — `---` fence pairs at head of file.
 7. **Mode markers** per [`role-mode-adherence`](../../rules/role-mode-adherence.md).
 
-Mangling any of these breaks Iron-Law surface host tool reads. Unit tests in `tests/test_condense_memory.py` lock each carve-out class as regression case.
+Mangling any of these breaks Iron-Law surface host tool reads. Unit tests in `tests/scripts/condense_memory.test.ts` lock each carve-out class as regression case.
 
 ## Idempotency contract — Step 9 guard
 
@@ -83,7 +83,7 @@ If you need to edit condensed memory file, run `--decondense` first, edit restor
 
 ## Sensitive-path gate
 
-Every read path passes through [`scripts/validate_safe_paths.py`](../../../scripts/validate_safe_paths.py) `assert_safe()` before bytes leave disk. Gate is security floor for Phase 2 (input-side condensation) per `step-16-telegraph-substance.md` Phase 0; rollback of gate is rollback of this skill.
+Every read path passes through [`src/scripts/validate_safe_paths.ts`](../../../src/scripts/validate_safe_paths.ts) `assert_safe()` before bytes leave disk. Gate is security floor for Phase 2 (input-side condensation) per `step-16-telegraph-substance.md` Phase 0; rollback of gate is rollback of this skill.
 
 CLI exit codes:
 
@@ -106,7 +106,7 @@ CLI exit codes:
 No published `telegraph-v2` baseline for input-side savings yet (Step 11 of `step-16-telegraph-substance.md` ships that). Until then, maintainer judges per-target whether condensation pays its readability cost. Suggested workflow:
 
 1. `wc -c <path>` before — record baseline char count.
-2. `python3 scripts/condense_memory.py <path>` — condense + back up.
+2. `./scripts-run src/scripts/condense_memory <path>` — condense + back up.
 3. `wc -c <path>` after — record post-condensation char count.
 4. Eyeball diff: does prose stay legible? Are all Iron-Law fences intact?
 5. If yes → commit both `<path>` and `<path>.original.md`. If no → `--decondense`.
@@ -116,8 +116,8 @@ Future `telegraph-v2.md` will tabulate realised input-token saving against `agen
 ## Cross-references
 
 - [`telegraph-speak`](../../rules/telegraph-speak.md) — runtime rule script mirrors for input-side targets; `telegraph.speak_scope` does **not** gate this script (input-side runs regardless).
-- [`scripts/validate_safe_paths.py`](../../../scripts/validate_safe_paths.py) — Phase 0 gate; ported from upstream Telegraph `63a91ec`.
-- [`scripts/condense_memory.py`](../../../scripts/condense_memory.py) — implementation.
-- [`tests/test_condense_memory.py`](../../../tests/test_condense_memory.py) — regression locks for each carve-out + idempotency + denylist.
+- [`src/scripts/validate_safe_paths.ts`](../../../src/scripts/validate_safe_paths.ts) — Phase 0 gate; ported from upstream Telegraph `63a91ec`.
+- [`src/scripts/condense_memory.ts`](../../../src/scripts/condense_memory.ts) — implementation.
+- [`tests/scripts/condense_memory.test.ts`](../../../tests/scripts/condense_memory.test.ts) — regression locks for each carve-out + idempotency + denylist.
 - [`docs/contracts/condensation-default-kill-criterion.md`](../../../docs/contracts/condensation-default-kill-criterion.md) — v1 verdict (output-side; informs but does not gate this skill).
 - [`agents-md-thin-root`](../agents-md-thin-root/SKILL.md) — caps consumer-shipped `templates/AGENTS.md`; this skill is one tool to land under cap.

package/dist/agent-src/skills/context-document/SKILL.md

@@ -72,6 +72,7 @@ agents/settings/contexts/                                # Project-wide contexts
 | **Knowledge card** | Trust-tiered cache of *expensive* (remote) structural evidence — negative facts + pointers durable, positive structure a hypothesis | `lodash.md`, `stripe-api.md` |
 | **Standards card (Class A)** | Coding standards **derived from real tooling config** — pointer + digest, `trust: high (config-derived)`, auto-refreshed when config changes; never a flattened claim | `coding-standards.md` (points at `ruff.toml`, `.editorconfig`) |
 | **Lesson card (Class C)** | Learned lesson — `symptom` (fact) split from `hypothesis` (decaying theory), `trust: low`, subject-not-person, anti-calcification decay; promoted only via human gate (accumulation layer is eval-gated) | `agents/memory/curated/lessons/<slug>.md` |
+| **Knowledge page (typed)** | Team-shared, lifecycle-typed knowledge grown while working the project — episodic sessions, semantic concepts, procedures on the way to a skill, small decisions | `agents/knowledge/concepts/api-response-shape.md` |
 
 ## Where to store contexts
 
@@ -81,6 +82,7 @@ agents/settings/contexts/                                # Project-wide contexts
 | Project-wide or cross-module | `agents/settings/contexts/` |
 | Module-specific | `{module_root}/{Module}/{agent_folder}/settings/contexts/` (resolved via `modules.root_paths` + `modules.agent_folder`; Laravel example: `app/Modules/{Module}/agents/settings/contexts/`) |
 | **Knowledge card** (committed) | `agents/knowledge/<source>.md` — fill from the `knowledge-card` template |
+| **Knowledge page (typed, committed)** | `agents/knowledge/{sessions,concepts,procedures,decisions}/<slug>.md` — fill from the `knowledge-pages` template |
 | **Evidence Report / probe dumps / absence log** (ephemeral) | `agents/memory/knowledge/session/` (gitignored, overwritten each task) |
 | If unsure | Ask the user |
 
@@ -112,6 +114,21 @@ pointer, trust tagging, multi-evidence consistency). See
 card-worthy and [`evidence-discipline`](../../agent-src/contexts/execution/evidence-discipline.md)
 for the full model.
 
+### Knowledge pages — the typed sibling directories
+
+Alongside knowledge cards, `agents/knowledge/{sessions,concepts,procedures,decisions}/`
+holds team-shared knowledge grown **while working the project** — coding
+standards observed, module structure, API endpoint shapes, recurring
+mistakes, small decisions not big enough for an ADR. Fill new pages from the
+[`knowledge-pages`](../../agent-src/templates/contexts/knowledge-pages.md)
+template.
+
+**Retrieval protocol — index-first, then grep, then read.** Read
+`agents/knowledge/INDEX.md` first (one line per page, regenerated by
+`src/scripts/generate_knowledge_index.ts`); `grep` keywords second; read the
+specific file third. Never enumerate every knowledge file directly — no
+search infrastructure by design.
+
 ### Shared vs. project-specific contexts
 
 **`.augment/contexts/`** — Part of the shared package. Describes the agent infrastructure:

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

@@ -63,16 +63,16 @@ Scripts resolve all paths **skill-relative** (per
 — they work from any cwd:
 
 ```bash
-python3 <skills-root>/corpus-grounding/scripts/ground.py search \
+./scripts-run <skills-root>/corpus-grounding/scripts/ground search \
   --manifest <skills-root>/<domain-skill>/data/manifest.json \
   "fintech dashboard" [--domain style] [--stack react] \
   [--filter "Severity=HIGH"] [--max-results 3] [--json]
 
-python3 <skills-root>/corpus-grounding/scripts/ground.py ground \
+./scripts-run <skills-root>/corpus-grounding/scripts/ground ground \
   --manifest <skills-root>/<domain-skill>/data/manifest.json \
   "luxury e-commerce" [--context '{"data_heavy": true}'] [--persist DIR]
 
-python3 <skills-root>/corpus-grounding/scripts/ground.py validate --manifest …
+./scripts-run <skills-root>/corpus-grounding/scripts/ground validate --manifest …
 ```
 
 `<skills-root>` is wherever skills are deployed (`~/.claude/skills/` for

package/dist/agent-src/skills/decision-record/SKILL.md

@@ -105,6 +105,12 @@ Write:
 - **Consequences:** what becomes true, what becomes harder, what
   becomes impossible. At least one of each, or admit that nothing
   becomes harder (rare; verify).
+- **Revisit-if:** the condition that would reopen this decision — new
+  evidence class, a changed constraint the decision assumed, or an
+  age threshold. A lock recorded without this line reads as permanent
+  law instead of a decision under current conditions; see
+  [`decision-revisit-gate`](../../rules/decision-revisit-gate.md) for
+  what happens when a later change is blocked by it without one.
 
 ### 5. Wire the supersession chain (if any)
 

package/dist/agent-src/skills/decision-review/SKILL.md

@@ -2,17 +2,14 @@
 model_tier: high
 name: decision-review
 description: "Use to audit a past architectural decision — did the chosen option hold up, what assumptions drifted, should the ADR be superseded? Backward review only; does not lock new choices."
-source: package
 domain: process
 workspaces:
   - engineering
 packs:
   - analysis-workbench
-lifecycle: active
 trust:
   level: professional
   confidence: medium
-  human_review_required: false
 install:
   default: false
   removable: true
@@ -40,6 +37,11 @@ install:
   have contributed to problems.
 - Preparation for a supersession: confirm the old decision actually
   needs replacing before writing a new ADR.
+- Routed here by [`decision-revisit-gate`](../../rules/decision-revisit-gate.md)
+  when a beneficial change is blocked by a recorded lock (honest-null
+  verdict, "don't relitigate" memory, budget canon, or ADR) — this skill
+  supplies the backward-audit procedure that rule's re-evaluation offer
+  points to.
 
 Do NOT use when:
 

package/dist/agent-src/skills/deep-reading-analyst/SKILL.md

@@ -170,13 +170,16 @@ Regardless of goal, close with three fixed blocks:
 ## Do NOT
 
 - Do NOT force-apply all frameworks at the user's chosen depth — drop ones that add no insight.
-- Do NOT copy text verbatim from the source — always reword for the user's understanding.
+- Do NOT copy text verbatim from the source — always reword for the user's understanding; any exact quote follows `content-quoting-floor` (≤15 words, one quote per source).
+- Do NOT restate a fresh-lookup-class claim (current status, version, price, date) from model memory — verify against a live source or flag it as unverified, per `asking-and-brevity-examples § Volatile-fact freshness`.
 - Do NOT use academic jargon without one-line definitions in the "key concepts" block.
 - Do NOT skip Step 5 — the takeaways + quick win are the load-bearing output, not optional decoration.
 - Do NOT route code reviews, diff stress-tests, or incident debugging through this skill.
 
 ## Reference modules
 
+All 8 framework modules are project-local guidelines under `docs/guidelines/agent-infra/` (full text adopted under the Reference-Guideline Sunset Policy). Each carries an `## ADOPT citation` footer pinning its MIT upstream origin.
+
 - [`scqa-framework`](../../../docs/guidelines/agent-infra/scqa-framework.md) — situation / complication / question / answer decomposition.
 - [`5w2h-analysis`](../../../docs/guidelines/agent-infra/5w2h-analysis.md) — completeness check (7 questions).
 - [`critical-thinking`](../../../docs/guidelines/agent-infra/critical-thinking.md) — argument quality / fallacy detection.
@@ -186,3 +189,5 @@ Regardless of goal, close with three fixed blocks:
 - [`systems-thinking`](../../../docs/guidelines/agent-infra/systems-thinking.md) — feedback loops + leverage points.
 - [`six-hats`](../../../docs/guidelines/agent-infra/six-hats.md) — White / Red / Black / Yellow / Green / Blue protocol.
 - [`comparison-matrix`](../../../docs/guidelines/agent-infra/comparison-matrix.md) — cross-source synthesis (L4 only).
+
+> **Deprecated — on-demand upstream loading.** Earlier revisions of this skill described loading some modules "on demand from a SHA-pinned URL", backed by an `external_source` frontmatter pin. That capability was **never wired** (no fetcher ever consumed the pin; it was schema-only metadata), so the pin and its `refresh_trigger` / `sunset_criterion` bundle have been **removed** — every module is local, no runtime fetch is needed. If a future revision genuinely needs live upstream content, build an explicit fetch step; do not reintroduce a non-functional metadata pin.

package/dist/agent-src/skills/design-intelligence/SKILL.md

@@ -169,7 +169,7 @@ corpus. Canon is a gap-filler, never an override of a registered brand value.
    product → style → color → landing → typography with decision rules):
 
    ```bash
-   python3 <skills-root>/corpus-grounding/scripts/ground.py ground \
+   ./scripts-run <skills-root>/corpus-grounding/scripts/ground ground \
      --manifest <skills-root>/design-intelligence/data/manifest.json \
      "<product type + mood + platform>" --json
    ```
@@ -235,7 +235,7 @@ finding classes instead of ad-hoc judgment:
 Per-framework Do/Don't corpora (16 stacks) ride the same manifest:
 
 ```bash
-python3 <skills-root>/corpus-grounding/scripts/ground.py search \
+./scripts-run <skills-root>/corpus-grounding/scripts/ground search \
   --manifest <skills-root>/design-intelligence/data/manifest.json \
   --stack react "list rerender memo" [--filter "Severity=HIGH"]
 ```
@@ -276,6 +276,9 @@ guidance + docs URLs from here instead of memory.
 | Orchestration gates + locks | `directives/ui/{design,review,polish}.py` — never import the engine |
 | WCAG audit method | [`accessibility-auditor`](../accessibility-auditor/SKILL.md) |
 | Token authoring | [`design-tokens`](../design-tokens/SKILL.md) |
+| Lo-fi structure exploration (pre-selection) | [`wireframe`](../wireframe/SKILL.md) — disposable greyscale variants |
+| Multiple hi-fi options (post-selection) | [`design-variations`](../design-variations/SKILL.md) — grounds each variation via this skill |
+| Fixed-canvas slide decks | [`html-deck`](../html-deck/SKILL.md) — own medium, still corpus-grounded |
 
 ## Gotchas
 

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

@@ -36,7 +36,7 @@
 ## How to use
 
 1. Select the style through the grounding engine first:
-   `python3 <skills-root>/corpus-grounding/scripts/ground.py search
+   `./scripts-run <skills-root>/corpus-grounding/scripts/ground search
    --manifest <skills-root>/design-intelligence/data/manifest.json
    --domain style "<product + mood>"`.
 2. Open the matching spec file and lift its token system (colors,

package/dist/agent-src/skills/design-tokens/SKILL.md

@@ -63,17 +63,17 @@ under `dark.semantic.*` overriding the same semantic names (emitted as a
 
 ```bash
 # tokens.json → CSS variables (primitives + semantic + components + .dark)
-python3 <skills-root>/design-tokens/scripts/tokens.py generate \
+./scripts-run <skills-root>/design-tokens/scripts/tokens generate \
   --config tokens.json -o assets/design-tokens.css
 
 # tokens.json → Tailwind theme.extend.colors snippet
-python3 …/tokens.py generate --config tokens.json --format tailwind
+./scripts-run …/tokens generate --config tokens.json --format tailwind
 
 # token-discipline lint: hardcoded hex/rgb/px/rem outside token files
-python3 …/tokens.py validate --dir src/ [--json]
+./scripts-run …/tokens validate --dir src/ [--json]
 
 # embeddable inline CSS for standalone HTML artifacts
-python3 …/tokens.py embed --tokens assets/design-tokens.css --minimal --style
+./scripts-run …/tokens embed --tokens assets/design-tokens.css --minimal --style
 ```
 
 `validate --json` emits findings with `"kind": "token_violation"` — the
@@ -92,6 +92,12 @@ four-operation split assigns to *rules/linters*, not the corpus).
    actually uses.
 2. **Ground the values** — `design-intelligence` query gives the
    WCAG-checked semantic color set + typography pairing for the product.
+   No brand and no corpus match (a genuinely from-scratch palette) →
+   derive the accent set in `oklch()` with shared lightness and chroma,
+   varied hue (`oklch(50% 0.15 250)` / `… 200` / `… 280`) — equal L/C
+   keeps the hues perceptually balanced where random hex codes drift in
+   saturation; cross-check the result against C1/C5 avoidance before
+   locking it in.
 3. **Author `tokens.json`** from the starter: fill primitives, point
    semantics at them, add `dark.semantic` overrides.
 4. **Generate** CSS vars (and the Tailwind snippet when the stack is

package/dist/agent-src/skills/design-variations/SKILL.md

@@ -0,0 +1,127 @@
+---
+model_tier: inherit
+name: design-variations
+description: "Produce 3+ substantively distinct hi-fi design variations — basic to bold, one file with tweak controls — when the user asks for options, alternatives, or \"show me a few takes\"."
+domain: product
+personas: []
+workspaces:
+  - engineering
+packs:
+  - frontend-design
+trust:
+  level: professional
+install:
+  removable: true
+execution:
+  type: manual
+---
+
+# Design Variations
+
+Produce multiple substantively distinct design variations so the user can
+reject what they don't want and combine what they do. One design is one bet;
+three variations surface preferences the user didn't know they had.
+
+## When to use
+
+- The user asks for options, alternatives, "different takes", "show me a few"
+  on hi-fi UI work (screen, component, flow, visual treatment).
+- After a `wireframe` round settled the structure and the user wants styled
+  directions on top of it.
+
+**When NOT to use:** structure/flow exploration before any styling decision →
+[`wireframe`](../wireframe/SKILL.md) (lo-fi, disposable). Selecting ONE
+grounded direction without generating alternatives →
+[`design-intelligence`](../design-intelligence/SKILL.md). Reviewing an
+existing design → [`design-review`](../design-review/SKILL.md).
+
+## Procedure
+
+1. **Baseline.** Confirm what is varied (screen / component / flow / visual
+   treatment), the existing design context (DESIGN.md, brand tokens, UI kit —
+   run [`existing-ui-audit`](../existing-ui-audit/SKILL.md) when the output
+   lands in a real codebase, per the `ui-audit-gate` rule), the count
+   (default 3, ceiling 6), and the user's priority axis.
+2. **Pick 2–4 axes** to vary across: visual treatment (tone, density, shadow,
+   radius, type weight), layout (centered/asymmetric, single/multi-column,
+   full-bleed/inset), interaction model (single page vs multi-step, modal vs
+   inline), hierarchy (what is primary), tone (minimal/formal/playful/
+   editorial), component style. Map axes onto the project's Taste Dials
+   (Variance/Motion/Density) where DESIGN.md defines them.
+3. **Spec each variation BEFORE building** — one line per variation naming
+   its distinct palette family, type pairing, and layout skeleton. Left
+   unspecified, variations converge on one default look; variety must be
+   designed, not hoped for. Ground palette/type picks via
+   [`design-intelligence`](../design-intelligence/SKILL.md) and emit values
+   as tokens per [`design-tokens`](../design-tokens/SKILL.md).
+4. **Build basic → bold.** Variation 1 = by-the-book (matches existing
+   patterns); variation 2 = refined (same structure, one or two dimensions
+   pushed — often the actual pick); variation 3 = novel (genuinely different
+   layout/metaphor/aesthetic, deliberately off-distribution). Cover both
+   ends — an all-safe set wastes the round, an all-wild set ignores the brief.
+5. **Substantive, not cosmetic.** Two variations that differ only in button
+   color or shadow opacity are one variation — drop and replace. Litmus: the
+   user can articulate the difference between any two variations in one
+   sentence.
+6. **Present in a single file** with tweak controls — never `v1.html` /
+   `v2.html` / `v3.html`. Structure-sharing variants toggle via CSS custom
+   properties + a small floating "Tweaks" panel (3–8 controls: color picker,
+   font/variant dropdown, density slider, section toggles, copy inputs);
+   persist chosen values in `localStorage`. Structurally distinct variants
+   render side-by-side with labels. Even unasked, expose 1–2 tweak axes by
+   default.
+7. **Caption each variation** (1–2 sentences naming the axis it flexes) and
+   **close with a recommendation** — a designer offers an opinion; the user
+   decides.
+8. **Quality gate.** Novel ≠ sloppy: run `lint_design_slop` over the emitted
+   file (every variation individually passes; a DESIGN.md-declared direction
+   suppresses its flags) and honor the outward-artifact hygiene floor (Q13 in
+   [`design-antipatterns.md`](../../../docs/guidelines/design-antipatterns.md)) —
+   no workspace paths, skill/tool names, or generator traces in the markup.
+   Final pass before delivery → [`design-review`](../design-review/SKILL.md).
+
+## Output format
+
+1. **One HTML/JSX file** containing all variations — side-by-side sections or
+   custom-property-driven variants with a floating Tweaks panel (3–8 controls,
+   `localStorage`-persisted); no per-variation file scatter.
+2. **Variation spec block** (comment or intro): per variation one line —
+   axis flexed, palette family, type pairing, layout skeleton.
+3. **Captions + one recommendation** naming the suggested pick and why, with
+   the trade-off of the runner-up.
+4. **Clean slop scan** — `lint_design_slop` output for the file (flags
+   resolved or DESIGN.md-justified), no Q13 violations.
+
+## Do NOT
+
+- Do NOT produce variations that differ only cosmetically (color swap, shadow
+  tweak) — replace with a substantive alternative.
+- Do NOT scatter `v1/v2/v3` files — one file, toggleable.
+- Do NOT let variations drift into one house style — spec palette/type/layout
+  per variation before building.
+- Do NOT skip the recommendation — presenting options without an opinion
+  pushes the decision cost back onto the user.
+- Do NOT invent brand values — consumer brand tokens win
+  (`brand-source-of-truth`); the corpus fills gaps only.
+- Do NOT embed system internals (paths, skill/tool names, generator traces)
+  in the emitted markup — Q13.
+
+## Gotcha
+
+- The "novel" variation is where slop concentrates: off-distribution choices
+  still pass the antipatterns catalog — deliberate ≠ default. Declare the
+  direction in the spec block so `lint_design_slop` gating reads intent.
+- Tweak panels leak into delivery: the panel must be removable (single
+  `<script>`/`<aside>` block, clearly marked) so the chosen variant ships
+  without exploration chrome.
+- More than 6 variations degrades choice quality — the user can't hold them
+  in mind; split into two rounds on different axes instead.
+
+## See also
+
+- [`wireframe`](../wireframe/SKILL.md) — lo-fi structure exploration before this.
+- [`design-intelligence`](../design-intelligence/SKILL.md) — grounded direction/palette/type selection.
+- [`design-tokens`](../design-tokens/SKILL.md) — token emission for variation values.
+- [`fe-design`](../fe-design/SKILL.md) — production heuristics the by-the-book variation follows.
+- [`design-review`](../design-review/SKILL.md) — final gate before delivery.
+- [`docs/guidelines/design-antipatterns.md`](../../../docs/guidelines/design-antipatterns.md) — catalog + Q13 outward-artifact hygiene.

package/dist/agent-src/skills/design-variations/evals/triggers.json

@@ -0,0 +1,17 @@
+{
+  "skill": "design-variations",
+  "last_eval": "2026-07-06",
+  "description": "5 should-trigger + 5 should-not-trigger. Should-trigger covers hi-fi options/alternatives requests. Should-not-trigger covers near-miss neighbors: wireframe (lo-fi structure exploration), design-intelligence (single grounded direction), design-review (review existing), image generation, and single-implementation feature work.",
+  "queries": [
+    {"q": "show me 3 different takes on this pricing page", "trigger": true},
+    {"q": "give me a few design options for the dashboard header", "trigger": true},
+    {"q": "I can't decide on the hero layout — build some alternatives I can compare", "trigger": true},
+    {"q": "make this settings screen in a safe version and a bold version", "trigger": true},
+    {"q": "let me play with color and density options on this card grid", "trigger": true},
+    {"q": "sketch a few rough layout directions for the signup flow before we style anything", "trigger": false, "note": "lo-fi structure exploration → wireframe"},
+    {"q": "pick a visual direction and palette for our new SaaS app", "trigger": false, "note": "single grounded direction → design-intelligence"},
+    {"q": "review this landing page design for issues", "trigger": false, "note": "review of existing design → design-review"},
+    {"q": "generate 3 logo images in Midjourney", "trigger": false, "note": "AI image generation → image-provider-routing / logo-generation"},
+    {"q": "build the checkout form component", "trigger": false, "note": "single implementation, no options ask → fe-design / stack executor"}
+  ]
+}

package/dist/agent-src/skills/emit-tickets/SKILL.md

@@ -40,7 +40,7 @@ Steps at a glance:
 5. Write `manifest.yml` with acyclic `dependency_graph` and empty `linear_state`.
 6. Append `<!-- ticket: T-NNN -->` markers back into the roadmap.
 7. (Re)generate `agents/tickets/_registry.yml`.
-8. Run `lint_ticket_buildable.py` until exit 0 (max 3 attempts).
+8. Run `lint_ticket_buildable` until exit 0 (max 3 attempts).
 9. Tell the user the bundle is ready: paste a ticket into Linear/Jira, or have the agent create it via a tracker MCP server (ADR-102 — no API export).
 
 ### §0 — Locate the roadmap
@@ -97,7 +97,7 @@ Apply this floor **before** marking any ticket `lite`:
 - Any miss → split the ticket further, or escalate to `medium`
 - Architectural / multi-module → `high` (use `adr-create` first)
 
-**Gate:** run `python3 src/scripts/lint_ticket_buildable.ts agents/tickets/{slug}/`
+**Gate:** run `./scripts-run src/scripts/lint_ticket_buildable agents/tickets/{slug}/`
 and fix all errors before declaring any ticket `lite`. This is the only
 authoritative check — do not skip it.
 
@@ -154,7 +154,7 @@ If the file does not exist, create it with the `bundles:` root.
 ### §8 — Run the buildability gate
 
 ```bash
-python3 src/scripts/lint_ticket_buildable.ts agents/tickets/{slug}/
+./scripts-run src/scripts/lint_ticket_buildable agents/tickets/{slug}/
 ```
 
 Fix every reported error. Re-run until exit code 0. Max 3 attempts; if still
@@ -181,7 +181,7 @@ Do NOT auto-create tracker issues — paste/MCP is the user's call.
 
 - Implement or invent ticket content not traceable to a roadmap step.
 - Touch `src/scripts/**` — orchestrate; do not reimplement.
-- Mark a ticket `lite` before `lint_ticket_buildable.py` passes.
+- Mark a ticket `lite` before `lint_ticket_buildable` passes.
 - Create a circular `dependency_graph` — verify acyclicity before writing.
 - Auto-create tracker issues — paste/MCP is the user's call (ADR-102, no API export).
 - Reuse T-NNN numbers from another bundle in the same registry.

package/dist/agent-src/skills/fe-design/SKILL.md

@@ -233,6 +233,26 @@ Animate `transform` and `opacity` only. Why: these run on the GPU compositor thr
 5. **Loading states** — Skeleton screens preferred over spinners. **Why:** skeletons reduce perceived wait time by showing structure immediately.
 6. **Error recovery** — Clear error messages with suggested actions. **Why:** "something went wrong" gives the user nothing to do; a specific message with a next step reduces support burden.
 
+## Craft details — typography & imagery
+
+- **`text-wrap: pretty` on body copy** — avoids widows/orphans without manual
+  breaks; cheap, no layout risk. (Display headlines: `text-wrap: balance`.)
+- **Honest placeholders over weak assets** — when the real image/illustration
+  is missing, use a striped placeholder with a monospace size label
+  (`repeating-linear-gradient(45deg, #E5E5E5 0 10px, #F5F5F5 10px 20px)` +
+  `product shot (1200×800)`) instead of a hand-drawn SVG or stock-ish filler.
+  A placeholder signals "asset needed"; a weak illustration signals nobody
+  had the asset. Never Lorem-ipsum copy in review-bound output
+  (per `output-discipline`) — use short label-style skeleton copy.
+
+## Presenting variants
+
+This skill produces ONE refined solution. When the user wants multiple options
+to compare (or a live tweak panel over one design), hand off to
+[`design-variations`](../design-variations/SKILL.md) — it owns the
+basic→bold variation method and the single-file tweak-panel mechanics
+(CSS custom properties + floating "Tweaks" panel + `localStorage`).
+
 ## Cross-task design memory — read DESIGN.md / PRODUCT.md first
 
 Before applying any heuristics from this reference, check the project root

package/dist/agent-src/skills/html-deck/SKILL.md

@@ -0,0 +1,123 @@
+---
+model_tier: inherit
+name: html-deck
+description: "Build a slide presentation as one HTML file — fixed 1920×1080 canvas letterboxed to any viewport, layout-system-first, type floors. Use for deck, slides, presentation, or pitch requests."
+domain: product
+personas: []
+workspaces:
+  - engineering
+packs:
+  - frontend-design
+trust:
+  level: professional
+install:
+  removable: true
+execution:
+  type: manual
+---
+
+# HTML Deck
+
+Build a slide presentation as a single HTML file with fixed-size slides that
+letterbox to any viewport. A deck is fixed-canvas media (default 1920×1080,
+16:9) — a different medium from responsive UI and from static posters.
+
+## When to use
+
+- The user asks for a deck, presentation, slides, or a pitch as a web/HTML
+  deliverable.
+
+**When NOT to use:** static single-image visual art (poster, PNG/PDF
+one-off) → [`canvas-design`](../canvas-design/SKILL.md). Responsive product
+UI → [`fe-design`](../fe-design/SKILL.md) + stack executor. Multiple styled
+options of one screen → [`design-variations`](../design-variations/SKILL.md).
+
+## Procedure
+
+1. **Discovery.** Confirm: audience (engineers / executives / customers —
+   determines tone and density), aspect ratio (16:9 default), slide count vs
+   time budget (~1 slide/minute; most decks land at 8–15), tone, source
+   material (PRD/doc — read it before sketching), speaker notes (off unless
+   requested), and the brand / design system. No brand →
+   [`design-intelligence`](../design-intelligence/SKILL.md) direction first.
+   A fully specified ask ("5-slide deck for the all-hands from this PRD")
+   skips the question round.
+2. **Commit to a layout system before any slide.** 4–6 layout types (cover,
+   section header, content, quote/pull-out, comparison two-column, closing/
+   CTA); per layout: background, headline size/position, body area, footer
+   treatment. 1–2 background colors across the deck (section headers may
+   break to a third). Record the system as a comment block at the top of the
+   file.
+3. **Start from the deck-shell starter** (`templates/deck-shell.html`,
+   skill-local): fixed-canvas scaling + letterboxing, keyboard/tap
+   navigation, slide counter, `localStorage` slide-index persistence,
+   print-to-PDF CSS. Each slide is a direct child `<section>` with a
+   1-indexed `data-screen-label` ("01 Title") so the user can reference
+   slides by the counter they see.
+4. **Build slide-by-slide, reveal early.** Show the file at 1–2 slides —
+   don't perfect 15 in private. Per slide: one primary message; **type
+   floors on a 1080p canvas: body ≥24px (32px+ preferred), headlines
+   60–96px+**; tokens from the active brand/design system (precedence per
+   `brand-source-of-truth`, emission per
+   [`design-tokens`](../design-tokens/SKILL.md)); honest placeholders for
+   missing imagery (striped background + monospace size label — see
+   [`fe-design § Craft details`](../fe-design/SKILL.md)); cut filler slides
+   ("Why choose us?", "About this deck") and data that doesn't support the
+   slide's point.
+5. **Speaker notes only when requested** — as a toggleable overlay or
+   per-slide `<aside hidden>`, never printed.
+6. **Quality gate.** Run `lint_design_slop` over the file; honor the
+   outward-artifact hygiene floor (Q13 in
+   [`design-antipatterns.md`](../../../docs/guidelines/design-antipatterns.md))
+   — no workspace paths, skill/tool names, or generator traces in markup or
+   comments; final pass → [`design-review`](../design-review/SKILL.md).
+   Verify scaling + navigation + persistence once in a browser before
+   delivery.
+
+## Output format
+
+1. **One self-contained HTML file** built on the deck-shell: fixed-canvas
+   `<section>` slides with 1-indexed `data-screen-label`s, letterbox scaling,
+   keyboard/tap navigation, slide counter, `localStorage` persistence,
+   print-to-PDF CSS.
+2. **Layout-system comment block** at the top of the file naming the 4–6
+   layout types and the background-color budget.
+3. **Type-floor compliance** — body ≥24px, headlines 60–96px+ on the 1080p
+   canvas; deviations only with a stated reason.
+4. **Clean slop scan + Q13** — `lint_design_slop` findings resolved or
+   DESIGN.md-justified; no system internals in the artifact.
+
+## Do NOT
+
+- Do NOT hand-roll the scaling/navigation — start from the deck-shell
+  starter; re-derived viewport math drifts.
+- Do NOT use responsive units/breakpoints inside slides — a deck is
+  fixed-canvas; the shell scales the whole stage.
+- Do NOT drop body text below 24px on a 1080p canvas — projector legibility
+  is the floor, not a preference.
+- Do NOT pad with filler slides or invented stats — every slide earns its
+  place.
+- Do NOT introduce per-slide ad-hoc colors/spacing — the layout system from
+  step 2 is the budget.
+- Do NOT embed system internals (paths, skill/tool names, generator traces)
+  in the file — Q13.
+
+## Gotcha
+
+- `localStorage` persistence keys on `location.pathname` — two decks served
+  from the same path share a slide index; rename the file per deck or extend
+  the key when hosting several decks under one path.
+- Print-to-PDF uses the fixed canvas per page: browsers map it onto the
+  paper size — set the print dialog to landscape + "fit to page" or the
+  slides clip; test one print preview before shipping a PDF.
+- Emoji in slide headlines is the CP5 tell exactly as in product UI — the
+  deck medium doesn't exempt it.
+
+## See also
+
+- `templates/deck-shell.html` (skill-local) — the canonical shell; copy, don't re-derive.
+- [`canvas-design`](../canvas-design/SKILL.md) — static posters/one-frame art (not decks).
+- [`design-intelligence`](../design-intelligence/SKILL.md) — direction/palette/type grounding.
+- [`design-tokens`](../design-tokens/SKILL.md) / [`brand-to-tokens`](../brand-to-tokens/SKILL.md) — token precedence and emission.
+- [`design-review`](../design-review/SKILL.md) — final review gate.
+- [`docs/guidelines/design-antipatterns.md`](../../../docs/guidelines/design-antipatterns.md) — catalog, CP5, Q13.