@bendyline/gilde 0.1.50 → 0.1.52

package/data/craftbook-templates/it/item-intake/versions/1.0.2/test.json

@@ -31,9 +31,6 @@
     "worker": {
       "name": "Casper",
       "role": "Curator"
-    },
-    "craftbookParams": {
-      "workPath": "tasks/eval"
     }
   },
   "mocks": [],

package/data/craftbook-templates/it/item-intake/versions/1.0.3/craftbook.json

@@ -0,0 +1,276 @@
+{
+  "id": "item-intake",
+  "name": "Catalog an Item",
+  "description": "Take new arrivals from shoebox to catalog: identify each item from its photo notes, record it with condition and an honest value basis, and flag what needs a better photo or an expert eye — uncertain identification stays uncertain in the record instead of becoming confident fiction. Locks the record fields and the honesty rule FIRST, then catalogs each item, then reports confidence and follow-ups. Use for cataloging new items, collection intake, 'catalog these arrivals', and adding finds to the collection.\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 intake (curator) — photo notes read, record fields and honesty rule locked → gated on artifact `{{workPath}}/scope.md` (markdown-notes)\n2. Catalog the arrivals (curator) — identify each item from its photos, record it honestly → gated on artifact `{{workPath}}/catalog.md` (markdown-notes)\n3. Write the intake report (copywriter) — items added, identification confidence, follow-up worklist → gated on artifact `{{workPath}}/report.md` (markdown-report)\n\nThe gates never advance with an unmet criterion, and loop back to the\nowning phase to fix named gaps.\n",
+  "entryStepId": "scope",
+  "triggers": [
+    "catalog new items",
+    "catalog these arrivals",
+    "add items to the collection",
+    "collection intake",
+    "record the new finds"
+  ],
+  "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": "scope",
+      "name": "Scope the intake",
+      "description": "photo notes read, record fields and honesty rule locked",
+      "prompt": "Define the intake before recording anything. Step 1: read the photo intake notes waiting in inbox/ and the catalog store (collection.json) — the record fields in use and how existing items are shaped. Step 2: lock the record rules — every new item gets an id, name, category, condition, and its photo reference; a value estimate is welcome but only with its basis named; identification that is uncertain stays uncertain in the record ('pressing unverified'), never silently confident. Step 3: decide the follow-up buckets — needs a better photo, needs an expert eye, needs a document (receipt, certificate). Step 4: write an acceptance-criteria checklist ('every inbox item recorded exactly once', 'every record complete on the required fields', 'no identification stated above what the photos support', 'follow-ups named per item where earned'). Call write_task_note with the item count + rules + checklist and write {{workPath}}/scope.md.\n\nThe deliverable `{{workPath}}/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": "curator",
+      "advanceWhen": {
+        "file": "{{workPath}}/scope.md",
+        "minBytes": 1,
+        "sniff": "nonempty",
+        "artifact": true
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "{{workPath}}/scope.md",
+            "bytes": 120,
+            "artifact": true
+          },
+          {
+            "kind": "sniff",
+            "file": "{{workPath}}/scope.md",
+            "sniff": "nonempty",
+            "artifact": true
+          }
+        ],
+        "onReject": "scope",
+        "maxAttempts": 3
+      },
+      "next": "catalog",
+      "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": "catalog",
+      "name": "Catalog the arrivals",
+      "description": "identify each item from its photos, record it honestly",
+      "prompt": "Catalog each arrival against the locked rules. Step 1: for every photo note in inbox/, identify the item as far as the evidence carries — name, category, and condition from what the photos actually show. Step 2: record each item through the project's collection capabilities with id, name, category, condition, and photo reference; add a value estimate only where you can name its basis. Step 3: where a detail is unreadable or the identification is uncertain, record the uncertainty on the item itself and put it in the right follow-up bucket. Step 4: keep a per-item log in {{workPath}}/catalog.md — what you identified, how confident, and why. Call write_task_note with the recorded count and the follow-up count.\n\nThe deliverable `{{workPath}}/catalog.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": "curator",
+      "advanceWhen": {
+        "file": "{{workPath}}/catalog.md",
+        "minBytes": 1,
+        "sniff": "nonempty",
+        "artifact": true
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "{{workPath}}/catalog.md",
+            "bytes": 120,
+            "artifact": true
+          },
+          {
+            "kind": "sniff",
+            "file": "{{workPath}}/catalog.md",
+            "sniff": "nonempty",
+            "artifact": true
+          }
+        ],
+        "onReject": "catalog",
+        "maxAttempts": 3
+      },
+      "next": "report",
+      "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": "report",
+      "name": "Write the intake report",
+      "description": "items added, identification confidence, follow-up worklist",
+      "prompt": "Write the intake report from the catalog log. Step 1: open with the headline — items added this intake, out of how many that came in. Step 2: give a line per item — what it is, its condition, and your identification confidence in plain words. Step 3: list the follow-ups by bucket — better photo, expert eye, missing document — so the next session has a worklist. Step 4: flag anything whose value or identification deserves a second opinion before it is trusted. Keep it precise and unhurried — gallery-grade records are built one honest intake at a time. Write {{workPath}}/report.md and call write_task_note with the added count and the follow-up count.\n\nThe deliverable `{{workPath}}/report.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": "copywriter",
+      "advanceWhen": {
+        "file": "{{workPath}}/report.md",
+        "minBytes": 1,
+        "sniff": "nonempty",
+        "artifact": true
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "{{workPath}}/report.md",
+            "bytes": 1000,
+            "artifact": true
+          },
+          {
+            "kind": "contains",
+            "file": "{{workPath}}/report.md",
+            "pattern": "(?:^|\\n)#{1,3}\\s+\\S",
+            "flags": "i",
+            "label": "at least one markdown heading",
+            "artifact": true
+          }
+        ],
+        "onReject": "report",
+        "maxAttempts": 4
+      },
+      "next": "evaluate",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "git",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "artifact",
+        "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 {{workPath}}/report.md, {{workPath}}/catalog.md, and the collection store and verify every criterion from {{workPath}}/scope.md: every inbox item was recorded exactly once with the required fields complete, every identification stays within what the photo notes support (spot-check each against its inbox note), uncertainty is recorded on the item rather than smoothed over, and every earned follow-up is on the worklist. Write PASS/FAIL per criterion; on any failure, name the gap and loop back to catalog.\n\nThe deliverable lives in the project's artifacts drawer — open `{{workPath}}/report.md` with `read_artifact`, not `read_file`.\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: \"catalog\" })` 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": "{{workPath}}/report.md",
+          "artifact": true
+        }
+      ],
+      "next": "catalog",
+      "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": "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/it/item-intake/versions/1.0.3/test.json

@@ -0,0 +1,145 @@
+{
+  "schemaVersion": 1,
+  "title": "Catalog an Item smoke eval",
+  "objective": "Task-specific eval for the Catalog an Item craftbook: three vinyl arrivals with pre-extracted photo notes become complete catalog records, with the unreadable pressing recorded as uncertain rather than guessed.",
+  "tags": [
+    "data"
+  ],
+  "prompt": "three new records came in for the vinyl collection — photo notes are in inbox/. can you cataloge them?",
+  "setup": {
+    "projectName": "Collection Catalog — Vinyl Intake Eval",
+    "about": "Self-contained eval project for item-intake. Arrival photos have already been pre-extracted into plain-text photo notes under inbox/. The catalog store lives at collection.json in the workspace; the intake write-up goes to tasks/eval/report.md.",
+    "missionObjectives": "Use the Catalog an Item craftbook/template, identify each arrival only as far as the seeded photo notes support, and finish with an updated collection.json plus tasks/eval/report.md — no network calls, real credentials, or live services.",
+    "files": [
+      {
+        "path": "inbox/item-1.txt",
+        "content": "PHOTO INTAKE NOTE (pre-extracted from arrival photos item-1-front.jpg, item-1-label.jpg)\n\nSleeve front: 'Miles Davis - Kind of Blue', Columbia.\nLabel: six-eye Columbia design, mono marking visible.\nSleeve: edge wear, small split starting at the top seam.\nVinyl: glossy, light hairlines under angled light.\nPhotos on file: photos/arrivals/item-1-front.jpg, photos/arrivals/item-1-label.jpg\n"
+      },
+      {
+        "path": "inbox/item-2.txt",
+        "content": "PHOTO INTAKE NOTE (pre-extracted from arrival photos item-2-front.jpg)\n\nSleeve front: 'Joni Mitchell - Blue', Reprise.\nSleeve: near mint, sharp corners, no ring wear.\nVinyl: clean, strong shine.\nPhotos on file: photos/arrivals/item-2-front.jpg\n"
+      },
+      {
+        "path": "inbox/item-3.txt",
+        "content": "PHOTO INTAKE NOTE (pre-extracted from arrival photos item-3-front.jpg, item-3-runout.jpg)\n\nSleeve front: 'Dave Brubeck - Time Out', Columbia.\nRunout etching: blurry in the photo - the pressing markings can not be read.\nSleeve: good shape overall, ballpoint writing on the back.\nPhotos on file: photos/arrivals/item-3-front.jpg, photos/arrivals/item-3-runout.jpg\n"
+      },
+      {
+        "path": "collection.json",
+        "content": "[\n  { \"id\": \"v-001\", \"name\": \"John Coltrane - Blue Train\", \"category\": \"jazz-lp\", \"condition\": \"VG+\", \"photo\": \"photos/v-001.jpg\", \"estValueUsd\": 220 },\n  { \"id\": \"v-002\", \"name\": \"Nina Simone - I Put a Spell on You\", \"category\": \"vocal-jazz-lp\", \"condition\": \"VG\", \"photo\": \"photos/v-002.jpg\", \"estValueUsd\": 95 },\n  { \"id\": \"v-003\", \"name\": \"The Beatles - Abbey Road\", \"category\": \"rock-lp\", \"condition\": \"VG+\", \"photo\": \"photos/v-003.jpg\", \"estValueUsd\": 85 },\n  { \"id\": \"v-004\", \"name\": \"Fleetwood Mac - Rumours\", \"category\": \"rock-lp\", \"condition\": \"NM\", \"photo\": \"photos/v-004.jpg\", \"estValueUsd\": 60 },\n  { \"id\": \"v-005\", \"name\": \"Bill Evans - Waltz for Debby\", \"category\": \"jazz-lp\", \"condition\": \"VG\", \"photo\": \"photos/v-005.jpg\", \"estValueUsd\": 240 },\n  { \"id\": \"v-006\", \"name\": \"Stevie Wonder - Innervisions\", \"category\": \"soul-lp\", \"condition\": \"VG+\", \"photo\": \"photos/v-006.jpg\", \"estValueUsd\": 55 },\n  { \"id\": \"v-007\", \"name\": \"Dolly Parton - Jolene\", \"category\": \"country-lp\", \"condition\": \"VG\", \"photo\": \"photos/v-007.jpg\", \"estValueUsd\": 45 },\n  { \"id\": \"v-008\", \"name\": \"Herbie Hancock - Head Hunters\", \"category\": \"jazz-lp\", \"condition\": \"VG+\", \"photo\": \"photos/v-008.jpg\", \"estValueUsd\": 70 }\n]\n"
+      }
+    ],
+    "worker": {
+      "name": "Casper",
+      "role": "Curator"
+    }
+  },
+  "mocks": [],
+  "success": {
+    "summary": "The catalog gains three complete records grounded in the photo notes, the unreadable pressing stays uncertain in the record, and the report lists the earned follow-ups.",
+    "deliverables": [
+      {
+        "path": "tasks/eval/report.md",
+        "kind": "markdown-report",
+        "minBytes": 600,
+        "checks": [
+          {
+            "kind": "contains",
+            "file": "tasks/eval/report.md",
+            "pattern": "(?:^|\\n)#{1,3}\\s+\\S",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/report.md",
+            "pattern": "confiden|uncertain|not sure|unclear|unverified|unread",
+            "flags": "i",
+            "label": "identification confidence stated honestly"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/report.md",
+            "pattern": "photo|expert",
+            "flags": "i",
+            "label": "follow-up buckets present"
+          }
+        ]
+      }
+    ],
+    "checks": [
+      {
+        "kind": "recordSchema",
+        "file": "collection.json",
+        "fields": [
+          {
+            "name": "id",
+            "type": "nonempty"
+          },
+          {
+            "name": "name",
+            "type": "nonempty"
+          },
+          {
+            "name": "category",
+            "type": "nonempty"
+          },
+          {
+            "name": "condition",
+            "type": "nonempty"
+          },
+          {
+            "name": "photo",
+            "type": "nonempty"
+          }
+        ],
+        "minRows": 11
+      },
+      {
+        "kind": "contains",
+        "file": "collection.json",
+        "pattern": "Kind of Blue",
+        "label": "the Miles Davis arrival is on record"
+      },
+      {
+        "kind": "contains",
+        "file": "collection.json",
+        "pattern": "Joni Mitchell",
+        "label": "the Joni Mitchell arrival is on record"
+      },
+      {
+        "kind": "contains",
+        "file": "collection.json",
+        "pattern": "Time Out|Brubeck",
+        "label": "the Brubeck arrival is on record"
+      }
+    ]
+  },
+  "rubric": {
+    "artifact": {
+      "path": "tasks/eval/report.md",
+      "kind": "markdown"
+    },
+    "axes": [
+      {
+        "name": "identification",
+        "description": "Each item is identified exactly as far as the photo notes carry — details in the record trace to the notes."
+      },
+      {
+        "name": "honesty",
+        "description": "The unreadable pressing is recorded as uncertain, never upgraded to a confident claim."
+      },
+      {
+        "name": "completeness",
+        "description": "Every arrival gets a complete record with id, name, category, condition, and photo reference."
+      },
+      {
+        "name": "care",
+        "description": "Follow-ups (better photo, expert eye) are named per item where earned."
+      }
+    ]
+  },
+  "qualityFocus": [
+    "evidence-bounded identification",
+    "recorded uncertainty",
+    "complete catalog records"
+  ]
+}

package/data/craftbook-templates/ju/juice-pass/versions/1.0.2/test.json

@@ -19,9 +19,6 @@
     "worker": {
       "name": "Ties",
       "role": "Game developer"
-    },
-    "craftbookParams": {
-      "workPath": "tasks/eval"
     }
   },
   "mocks": [],