@bendyline/gilde 0.1.51 → 0.1.52

package/data/craftbook-templates/kp/kpi-scorecard/versions/1.0.3/craftbook.json

@@ -0,0 +1,285 @@
+{
+  "id": "kpi-scorecard",
+  "name": "KPI Scorecard Page",
+  "description": "Build a single-page KPI scorecard that rolls a dataset up into the handful of numbers a team tracks, each shown with its current value, target, and period-over-period change. Defines the KPIs and their exact formulas and targets FIRST (so a number is never ambiguous), then computes the rollups, then builds a clean self-contained HTML scorecard with KPI tiles, deltas, and a tiny trend sparkline each. Use this for a KPI dashboard, a metrics scorecard, an executive summary page, a North Star metrics view, or rolling raw data into headline numbers.\n\nA gallery craftbook generated from an archetype spec. It runs\n`phase → (per-phase gate) → … → evaluate → (loop) → finish`. Each build\nphase that produces a checkable artifact is followed by a **runtime\ngate-checkpoint** — the runtime verifies the artifact and routes with no\nmodel turn, looping back to redo the phase on a miss. The final `evaluate`\nstep holds a static deliverable gate plus a reviewer QA pass. What it adds\nover the generic `build-loop`: a specialist role per phase, a\ndomain-correct ordering, and a concrete per-phase quality bar.\n\nDeliverables marked \"artifact\" land in the project's artifacts drawer (`write_artifact` / `read_artifact`), not the shipped workspace — review output is not product source.\n\nPhases:\n\n1. Scope the KPIs (planner) — lock each KPI's formula, target, and comparison period → gated on artifact `{{workPath}}/kpi-scope.md` (markdown-notes)\n2. Compute the rollups (data-analyst) — compute each KPI's value, delta, and trend series → gated on `data/kpis.json` (json)\n3. Build the scorecard (developer) — render a self-contained KPI scorecard page → gated on `index.html` (html-page)\n\nThe gates never advance with an unmet criterion, and loop back to the\nowning phase to fix named gaps.\n",
+  "entryStepId": "kpi-scope",
+  "triggers": [
+    "build a kpi scorecard",
+    "metrics dashboard page",
+    "executive metrics summary",
+    "north star metrics",
+    "roll data into kpis"
+  ],
+  "paramSchema": {
+    "type": "object",
+    "properties": {
+      "workPath": {
+        "type": "string",
+        "title": "Working folder",
+        "description": "Per-task working folder in the artifacts drawer. Defaults to this task's own folder so runs never collide; override with a stable name when you deliberately want runs to share files.",
+        "default": "{{task.dir}}"
+      }
+    }
+  },
+  "steps": [
+    {
+      "id": "kpi-scope",
+      "name": "Scope the KPIs",
+      "description": "lock each KPI's formula, target, and comparison period",
+      "prompt": "Define each KPI precisely before computing anything. Step 1: list the 4-8 KPIs the scorecard shows, ordered by importance. Step 2: for each, write the exact formula (numerator / denominator), the unit and format (count, %, currency), and the time window. Step 3: define the comparison — period-over-period (this month vs last) or vs a target — and the target value where one exists. Step 4: define 'good' direction (up is good vs down is good) so deltas can be colored. Step 5: write an acceptance-criteria checklist ('every KPI has a precise formula and unit', 'each tile shows current value + delta vs comparison', 'targets shown where defined', 'good/bad direction is correct', 'a small trend is shown per KPI'). Call write_task_note and write the KPI definitions + targets + checklist to {{workPath}}/kpi-scope.md.\n\nThe deliverable `{{workPath}}/kpi-scope.md` lands in the project's artifacts drawer — write it with `write_artifact` and read it back with `read_artifact`; the shipped workspace stays untouched.",
+      "suggestedRole": "planner",
+      "advanceWhen": {
+        "file": "{{workPath}}/kpi-scope.md",
+        "minBytes": 1,
+        "sniff": "nonempty",
+        "artifact": true
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "{{workPath}}/kpi-scope.md",
+            "bytes": 120,
+            "artifact": true
+          },
+          {
+            "kind": "sniff",
+            "file": "{{workPath}}/kpi-scope.md",
+            "sniff": "nonempty",
+            "artifact": true
+          }
+        ],
+        "onReject": "kpi-scope",
+        "maxAttempts": 3
+      },
+      "next": "compute",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "artifact",
+        "additionalOutputMedia": [
+          "task-note"
+        ]
+      }
+    },
+    {
+      "id": "compute",
+      "name": "Compute the rollups",
+      "description": "compute each KPI's value, delta, and trend series",
+      "prompt": "Compute every scoped KPI from the data. Step 1: for each KPI apply its exact formula over the current window and record the value. Step 2: compute the same for the comparison window and derive the absolute and % delta and whether it's an improvement given the good-direction. Step 3: build a short trend series (e.g. last 6-12 periods) per KPI for the sparkline. Step 4: attach the target and the gap to target where defined. Step 5: emit data/kpis.json — per KPI { name, value, unit, delta, delta_pct, improved, target, trend[] }. Double-check a denominator isn't zero and a % isn't mis-scaled. Call write_task_note with the headline KPI values and deltas.",
+      "suggestedRole": "data-analyst",
+      "advanceWhen": {
+        "file": "data/kpis.json",
+        "minBytes": 1,
+        "sniff": "json-valid"
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "data/kpis.json",
+            "bytes": 2
+          }
+        ],
+        "scripts": [
+          {
+            "name": "checkJsonValid",
+            "scope": "standard",
+            "inputs": {
+              "file": "data/kpis.json"
+            }
+          }
+        ],
+        "onReject": "compute",
+        "maxAttempts": 3
+      },
+      "next": "build",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web"
+        ],
+        "outputMedium": "workspace",
+        "additionalOutputMedia": [
+          "task-note"
+        ]
+      }
+    },
+    {
+      "id": "build",
+      "name": "Build the scorecard",
+      "description": "render a self-contained KPI scorecard page",
+      "prompt": "Build a single self-contained index.html scorecard from data/kpis.json. Step 1: render one KPI tile per metric in priority order, each showing the formatted current value, the metric name, and the delta vs comparison with an up/down arrow colored by improvement (green good, red bad per the good-direction). Step 2: show the target and gap where defined. Step 3: draw a tiny inline sparkline (canvas/SVG) of the trend series in each tile — no external chart CDN that could fail offline. Step 4: lay the tiles out in a responsive grid, headline KPIs largest. Step 5: ensure values are real (from the JSON), formatting is correct, and there are no console errors. On a loop-back, fix only the named gaps. Call write_task_note with the path and KPI count.",
+      "suggestedRole": "developer",
+      "consumes": [
+        {
+          "file": "data/kpis.json"
+        }
+      ],
+      "advanceWhen": {
+        "file": "index.html",
+        "minBytes": 1,
+        "sniff": "html-complete"
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "index.html",
+            "bytes": 800
+          },
+          {
+            "kind": "htmlLint",
+            "file": "index.html"
+          }
+        ],
+        "scripts": [
+          {
+            "name": "checkHtmlComplete",
+            "scope": "standard",
+            "inputs": {
+              "file": "index.html"
+            }
+          }
+        ],
+        "onReject": "build",
+        "maxAttempts": 4
+      },
+      "next": "evaluate",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web"
+        ],
+        "outputMedium": "workspace",
+        "additionalOutputMedia": [
+          "task-note"
+        ]
+      }
+    },
+    {
+      "id": "evaluate",
+      "name": "Evaluate",
+      "description": "Grade the deliverable against every acceptance criterion. All pass → finish; any fail → loop back and fix the gap.",
+      "prompt": "Open index.html (render it if possible, else trace the source) and check against {{workPath}}/kpi-scope.md and data/kpis.json. Verify: every scoped KPI appears as a tile with its formatted current value; each tile shows a delta vs the comparison with the correct good/bad color direction; targets and gaps are shown where defined; a trend sparkline renders per KPI from real data; and there are no console errors. Re-derive one KPI's value and delta from the data. Write PASS/FAIL per criterion; any FAIL loops back to build.\n\nThen route — this is the whole point of the loop:\n\n- **Every criterion PASSES →** call `advance_task_step({ ref, stepId: \"evaluate\", next: \"finish\" })`.\n- **Any criterion FAILS →** write the specific gaps to notes, then call `advance_task_step({ ref, stepId: \"evaluate\", next: \"build\" })` to loop back. The builder fixes exactly those gaps.\n\nNever route to `finish` while any criterion is unmet. The build phase's completion gate already blocked a grossly-incomplete deliverable; your job is the judgment an automated check cannot make (does it actually work, read well, look right). After ~3 unproductive loops, stop and report DONE_WITH_CONCERNS so the user can step in.",
+      "suggestedRole": "reviewer",
+      "consumes": [
+        {
+          "file": "index.html"
+        }
+      ],
+      "next": "build",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "task-note"
+      }
+    },
+    {
+      "id": "finish",
+      "name": "Finish",
+      "description": "All acceptance criteria met. Stamp a short summary and report DONE.",
+      "prompt": "Every acceptance criterion passed. Write a one-paragraph DONE summary to task notes via `write_task_note`: what was built, the deliverable path(s), and a one-line confirmation that each criterion is met. Then report DONE.",
+      "suggestedRole": "developer",
+      "terminal": true,
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "task-note"
+      }
+    }
+  ],
+  "version": "1.0.3",
+  "releasedAt": "2026-09-01T20:00:00Z",
+  "minGezelVersion": "1.26233"
+}

package/data/craftbook-templates/kp/kpi-scorecard/versions/1.0.3/test.json

@@ -0,0 +1,97 @@
+{
+  "schemaVersion": 1,
+  "title": "KPI Scorecard Page smoke eval",
+  "objective": "Self-contained smoke eval for the KPI Scorecard Page craftbook using the html-page generic harness.",
+  "tags": [
+    "html-page"
+  ],
+  "prompt": "Can you biuld us a little page for this? The content notes are in source/page-content.md. Save it as index.html — one file, nothing fancy needed on our end.",
+  "setup": {
+    "projectName": "KPI Scorecard Page Eval",
+    "about": "Self-contained eval project for kpi-scorecard. Seeded inputs are under workspace/source or workspace/fixtures; final deliverable is workspace/index.html.",
+    "missionObjectives": "Use the KPI Scorecard Page craftbook/template, read the seeded local fixtures, and write index.html without network calls, real credentials, or live services.",
+    "files": [
+      {
+        "path": "source/brief.md",
+        "content": "# KPI Scorecard Page Eval Brief\n\nClient: Boreal Desk, a home-office accessories company.\nAudience: operations leads who need an artifact they can use this week.\n\nFixed source facts for grounding:\n- The returns desk pilot covered 18 SKUs.\n- Median first response improved from 18 hours to 6 hours.\n- Preventable refund leakage fell from 14.2% to 8.9%.\n- The top unresolved complaint is status silence after photo submission.\n- Required next actions are automated status emails, barcode-exception training, and a weekly Finance exception export.\n\nUse these facts when the task asks for prose, analysis, copy, UI content, or test data. Do not use live web services, real credentials, or current outside data.\n\nCraftbook under test: kpi-scorecard - KPI Scorecard Page.\n"
+      },
+      {
+        "path": "source/page-content.md",
+        "content": "# Boreal Desk Support Board\n\nBuild content around a support operations board. Required labels: Open issues, Status silence, Response time, Refund leakage, Next action. Required metric text: 18 hours to 6 hours and 14.2% to 8.9%.\n"
+      }
+    ],
+    "worker": {
+      "name": "Jules",
+      "role": "Developer"
+    }
+  },
+  "mocks": [],
+  "success": {
+    "summary": "index.html is a source-grounded self-contained HTML page/tool.",
+    "deliverables": [
+      {
+        "path": "index.html",
+        "kind": "html-page",
+        "minBytes": 3000,
+        "checks": [
+          {
+            "kind": "cssMinBytes",
+            "bytes": 650,
+            "file": "index.html"
+          },
+          {
+            "kind": "jsParses",
+            "file": "index.html"
+          },
+          {
+            "kind": "contains",
+            "file": "index.html",
+            "pattern": "Boreal|Support Board|returns",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "index.html",
+            "pattern": "18\\s*hours|6\\s*hours|14\\.2%|8\\.9%",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "index.html",
+            "pattern": "button|input|filter|tab|card|section|nav",
+            "flags": "i"
+          }
+        ]
+      }
+    ]
+  },
+  "rubric": {
+    "artifact": {
+      "path": "index.html",
+      "kind": "html"
+    },
+    "axes": [
+      {
+        "name": "structure",
+        "description": "Layout and hierarchy fit the content; sections are scannable and coherent."
+      },
+      {
+        "name": "grounding",
+        "description": "Seeded source content is used faithfully — no invented facts or dropped requirements."
+      },
+      {
+        "name": "interactivity",
+        "description": "Controls and states the page claims to offer actually work."
+      },
+      {
+        "name": "polish",
+        "description": "Styling is intentional and responsive rather than unstyled or broken."
+      }
+    ]
+  },
+  "qualityFocus": [
+    "self-contained UI",
+    "source grounding",
+    "responsive structure"
+  ]
+}

package/data/craftbook-templates/la/landing-copy/versions/1.0.3/craftbook.json

@@ -0,0 +1,275 @@
+{
+  "id": "landing-copy",
+  "name": "Landing Page Copy",
+  "description": "Write conversion-focused copy for a single landing page: it first locks the offer, the one primary audience, the single desired action, and the core promise, then drafts a full message hierarchy — hero headline and subhead, value props, social proof, objection handling, and one repeated CTA — then edits ruthlessly for clarity, specificity, and persuasive flow. Covers landing page copy, sales copy, conversion copywriting, hero copy, CTA copy, and value proposition writing. Offer-first ordering matters: a small model that nails the single offer and one CTA before writing avoids the diluted page that lists ten features, hedges the promise, and offers three competing actions.\n\nA gallery craftbook generated from an archetype spec. It runs\n`phase → (per-phase gate) → … → evaluate → (loop) → finish`. Each build\nphase that produces a checkable artifact is followed by a **runtime\ngate-checkpoint** — the runtime verifies the artifact and routes with no\nmodel turn, looping back to redo the phase on a miss. The final `evaluate`\nstep holds a static deliverable gate plus a reviewer QA pass. What it adds\nover the generic `build-loop`: a specialist role per phase, a\ndomain-correct ordering, and a concrete per-phase quality bar.\n\nDeliverables marked \"artifact\" land in the project's artifacts drawer (`write_artifact` / `read_artifact`), not the shipped workspace — review output is not product source.\n\nPhases:\n\n1. Offer scope (planner) — lock the offer, one audience, one CTA, and the core promise → gated on artifact `{{workPath}}/offer-scope.md` (markdown-notes)\n2. Draft the copy (copywriter) — write the full message hierarchy section by section → gated on `landing-copy.md` (markdown-doc)\n3. Editorial pass (reviewer) — sharpen specificity, persuasion, and a single clear CTA → gated on `landing-copy.md` (markdown-doc)\n\nThe gates never advance with an unmet criterion, and loop back to the\nowning phase to fix named gaps.\n",
+  "entryStepId": "offer-scope",
+  "triggers": [
+    "write landing page copy",
+    "conversion copy",
+    "sales page copy",
+    "write hero copy",
+    "value proposition"
+  ],
+  "paramSchema": {
+    "type": "object",
+    "properties": {
+      "workPath": {
+        "type": "string",
+        "title": "Working folder",
+        "description": "Per-task working folder in the artifacts drawer. Defaults to this task's own folder so runs never collide; override with a stable name when you deliberately want runs to share files.",
+        "default": "{{task.dir}}"
+      }
+    }
+  },
+  "steps": [
+    {
+      "id": "offer-scope",
+      "name": "Offer scope",
+      "description": "lock the offer, one audience, one CTA, and the core promise",
+      "prompt": "1) State the OFFER in one sentence: the product/service and what the visitor gets. 2) Define exactly ONE primary audience and the specific pain or desire that brought them. 3) Pick the ONE primary call-to-action (e.g. 'Start free trial', 'Book a demo') — secondary CTAs are out of scope. 4) Write the core promise/value proposition in one line ('the one thing'). 5) List 3-5 concrete benefits (outcomes, not features) and the proof for each (stat, testimonial, guarantee). 6) List the top 2-3 objections a visitor will have and the rebuttal for each. Write an acceptance-criteria checklist (e.g. 'one headline states a specific outcome, not a vague slogan'; 'benefits are outcomes with proof, not feature lists'; 'exactly one CTA, repeated'; 'every objection is addressed'; 'reading level grade <=8'). Use write_task_note for the offer brief + checklist, and write the same to the produces path. No page copy yet.\n\nThe deliverable `{{workPath}}/offer-scope.md` lands in the project's artifacts drawer — write it with `write_artifact` and read it back with `read_artifact`; the shipped workspace stays untouched.",
+      "suggestedRole": "planner",
+      "advanceWhen": {
+        "file": "{{workPath}}/offer-scope.md",
+        "minBytes": 1,
+        "sniff": "nonempty",
+        "artifact": true
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "{{workPath}}/offer-scope.md",
+            "bytes": 120,
+            "artifact": true
+          },
+          {
+            "kind": "sniff",
+            "file": "{{workPath}}/offer-scope.md",
+            "sniff": "nonempty",
+            "artifact": true
+          }
+        ],
+        "onReject": "offer-scope",
+        "maxAttempts": 3
+      },
+      "next": "draft",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "artifact",
+        "additionalOutputMedia": [
+          "task-note"
+        ]
+      }
+    },
+    {
+      "id": "draft",
+      "name": "Draft the copy",
+      "description": "write the full message hierarchy section by section",
+      "prompt": "Write the landing page copy as a structured document with clear section labels. 1) HERO: a specific, benefit-driven headline (no vague slogans), a supporting subhead, and the primary CTA button label. 2) VALUE PROPS: 3-5 benefit blocks, each a short header + 1-2 sentences, leading with the outcome. 3) SOCIAL PROOF: testimonial(s) or trust signals with the proof from the brief. 4) OBJECTION HANDLING: a short FAQ or reassurance block answering each objection. 5) FINAL CTA: restate the promise and repeat the single CTA. Write tight, concrete, second-person copy; cut adjectives that do not earn their place. Output markdown with the section labels as H2s. Save to the produces path and note which checklist items pass.",
+      "suggestedRole": "copywriter",
+      "advanceWhen": {
+        "file": "landing-copy.md",
+        "minBytes": 1,
+        "sniff": "nonempty"
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "landing-copy.md",
+            "bytes": 500
+          }
+        ],
+        "scripts": [
+          {
+            "name": "checkContains",
+            "scope": "standard",
+            "inputs": {
+              "file": "landing-copy.md",
+              "pattern": "(?:^|\\n)#{1,3}\\s+\\S",
+              "flags": "i"
+            }
+          }
+        ],
+        "onReject": "draft",
+        "maxAttempts": 3
+      },
+      "next": "edit",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web"
+        ],
+        "outputMedium": "workspace"
+      }
+    },
+    {
+      "id": "edit",
+      "name": "Editorial pass",
+      "description": "sharpen specificity, persuasion, and a single clear CTA",
+      "prompt": "Edit the copy in place against the checklist. 1) Test the headline: is it a specific outcome a competitor could not also claim? If it is generic, rewrite it. 2) Convert any feature-speak into outcomes; make every benefit concrete and, where possible, quantified. 3) Confirm there is exactly ONE call-to-action and it is repeated verbatim in hero and final sections. 4) Verify each objection from the brief is answered. 5) Cut hype words, hedges, and filler; tighten to a grade-8 reading level. Rewrite weak lines directly. Save the polished copy to the produces path and note any criteria still unmet.",
+      "suggestedRole": "reviewer",
+      "advanceWhen": {
+        "file": "landing-copy.md",
+        "minBytes": 1,
+        "sniff": "nonempty"
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "landing-copy.md",
+            "bytes": 500
+          }
+        ],
+        "scripts": [
+          {
+            "name": "checkContains",
+            "scope": "standard",
+            "inputs": {
+              "file": "landing-copy.md",
+              "pattern": "(?:^|\\n)#{1,3}\\s+\\S",
+              "flags": "i"
+            }
+          }
+        ],
+        "onReject": "edit",
+        "maxAttempts": 4
+      },
+      "next": "evaluate",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web"
+        ],
+        "outputMedium": "workspace"
+      }
+    },
+    {
+      "id": "evaluate",
+      "name": "Evaluate",
+      "description": "Grade the deliverable against every acceptance criterion. All pass → finish; any fail → loop back and fix the gap.",
+      "prompt": "Read landing-copy.md and grade against the checklist. Check EACH and write PASS/FAIL: (a) one specific, outcome-driven headline (not a vague slogan); (b) benefits are outcomes with proof, not raw feature lists; (c) exactly one CTA, repeated identically in hero and final; (d) every brief objection is addressed; (e) social proof present; (f) tight, concrete, grade-8 copy with no hype filler. Copy that reads nicely but offers competing CTAs or buries the promise FAILS.\n\nThen route — this is the whole point of the loop:\n\n- **Every criterion PASSES →** call `advance_task_step({ ref, stepId: \"evaluate\", next: \"finish\" })`.\n- **Any criterion FAILS →** write the specific gaps to notes, then call `advance_task_step({ ref, stepId: \"evaluate\", next: \"draft\" })` to loop back. The builder fixes exactly those gaps.\n\nNever route to `finish` while any criterion is unmet. The build phase's completion gate already blocked a grossly-incomplete deliverable; your job is the judgment an automated check cannot make (does it actually work, read well, look right). After ~3 unproductive loops, stop and report DONE_WITH_CONCERNS so the user can step in.",
+      "suggestedRole": "reviewer",
+      "consumes": [
+        {
+          "file": "landing-copy.md"
+        }
+      ],
+      "next": "draft",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "task-note"
+      }
+    },
+    {
+      "id": "finish",
+      "name": "Finish",
+      "description": "All acceptance criteria met. Stamp a short summary and report DONE.",
+      "prompt": "Every acceptance criterion passed. Write a one-paragraph DONE summary to task notes via `write_task_note`: what was built, the deliverable path(s), and a one-line confirmation that each criterion is met. Then report DONE.",
+      "suggestedRole": "developer",
+      "terminal": true,
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "task-note"
+      }
+    }
+  ],
+  "version": "1.0.3",
+  "releasedAt": "2026-09-01T20:00:00Z",
+  "minGezelVersion": "1.26233"
+}