@bendyline/gilde 0.1.50 → 0.1.52

package/data/craftbook-templates/ci/ci-pipeline/versions/2.0.1/test.json

@@ -0,0 +1,225 @@
+{
+  "schemaVersion": 1,
+  "title": "CI pipeline grounded in real scripts",
+  "objective": "The craftbook authors a CI workflow for a repo that has exactly two scripts (test and lint) and no typecheck. The oracle verifies a real workflow exists under .github/workflows/, runs both real scripts with checkout and node setup pinned to major versions, triggers on push and pull_request, and invents nothing — no typecheck stage, no publish job, no fabricated secrets.",
+  "tags": [
+    "code",
+    "ci",
+    "devops",
+    "tactical-fleet"
+  ],
+  "prompt": "Use the CI Pipeline from Scratch craftbook: this repo has no CI at all. Set up a pipeline that runs on every push and pull request using what the repo already has — do not add stages for tools the repo does not use. Standard GitHub Actions layout.",
+  "setup": {
+    "projectName": "Relay Notes",
+    "about": "A small Node library with two package scripts: `npm run test` (node --test) and `npm run lint` (a repo-local lint script). No typecheck, no build step, no publish configuration.",
+    "files": [
+      {
+        "path": "package.json",
+        "content": "{\n  \"name\": \"relay-notes\",\n  \"private\": true,\n  \"version\": \"1.0.0\",\n  \"type\": \"module\",\n  \"scripts\": {\n    \"test\": \"node --test\",\n    \"lint\": \"node tools/lint.mjs\"\n  }\n}\n"
+      },
+      {
+        "path": "package-lock.json",
+        "content": "{\n  \"name\": \"relay-notes\",\n  \"version\": \"1.0.0\",\n  \"lockfileVersion\": 3,\n  \"requires\": true,\n  \"packages\": {\n    \"\": {\n      \"name\": \"relay-notes\",\n      \"version\": \"1.0.0\"\n    }\n  }\n}\n"
+      },
+      {
+        "path": "src/greet.js",
+        "content": "export function greet(name) {\n  const label = String(name ?? '').trim();\n  if (label === '') return 'Hello there!';\n  return `Hello, ${label}!`;\n}\n"
+      },
+      {
+        "path": "tools/lint.mjs",
+        "content": "import { readFileSync, readdirSync } from 'node:fs';\n\nlet problems = 0;\nfor (const file of readdirSync('src')) {\n  if (!file.endsWith('.js')) continue;\n  const text = readFileSync(`src/${file}`, 'utf8');\n  if (/\\t/.test(text)) {\n    console.error(`lint: tab indentation in src/${file}`);\n    problems++;\n  }\n}\nif (problems > 0) process.exit(1);\nconsole.log('lint ok');\n"
+      },
+      {
+        "path": "tests/greet.test.mjs",
+        "content": "import assert from 'node:assert/strict';\nimport { test } from 'node:test';\nimport { greet } from '../src/greet.js';\n\ntest('greets a trimmed name', () => {\n  assert.equal(greet('  Mira '), 'Hello, Mira!');\n});\n\ntest('falls back for an empty name', () => {\n  assert.equal(greet(''), 'Hello there!');\n});\n"
+      },
+      {
+        "path": "tests/verify-ci.mjs",
+        "content": "import assert from 'node:assert/strict';\nimport { readFileSync, readdirSync } from 'node:fs';\nimport { join } from 'node:path';\n\nconst dir = '.github/workflows';\nlet names = [];\ntry {\n  names = readdirSync(dir).filter((f) => /\\.ya?ml$/i.test(f));\n} catch {}\nassert.ok(names.length >= 1, 'no workflow file exists under .github/workflows/');\nconst raw = names.map((f) => readFileSync(join(dir, f), 'utf8')).join('\\n---\\n');\n// Judge only active YAML: a comment honestly noting \"no typecheck script\"\n// must never trip the anti-invention assertions below.\nconst all = raw.replace(/#[^\\n]*/g, '');\nconsole.log('CIPIPE_ORACLE workflow found');\n\n// Checkout and Node setup, pinned to major versions.\nassert.ok(/actions\\/checkout@v\\d+/.test(all), 'uses actions/checkout pinned to a major version');\nassert.ok(/actions\\/setup-node@v\\d+/.test(all), 'uses actions/setup-node pinned to a major version');\nconsole.log('CIPIPE_ORACLE actions ok');\n\n// Runs both real scripts.\nassert.ok(/\\bnpm\\s+(?:run\\s+)?test\\b/.test(all), 'the workflow runs the test script');\nassert.ok(/\\bnpm\\s+run\\s+lint\\b/.test(all), 'the workflow runs the lint script');\nconsole.log('CIPIPE_ORACLE scripts ok');\n\n// Invokes nothing the repo does not have.\nassert.ok(!/typecheck/i.test(all), 'no invented typecheck stage - the repo has no such script');\nconst pkg = JSON.parse(readFileSync('package.json', 'utf8'));\nfor (const m of all.matchAll(/npm\\s+run\\s+([A-Za-z0-9:_-]+)/g)) {\n  if (m[1].startsWith('-')) continue;\n  assert.ok(pkg.scripts && pkg.scripts[m[1]], `workflow invokes \"npm run ${m[1]}\" but package.json has no such script`);\n}\nconsole.log('CIPIPE_ORACLE no invented commands');\n\n// Triggers on push and pull_request.\nassert.ok(/\\bpush\\b/.test(all), 'triggers on push');\nassert.ok(/\\bpull_request\\b/.test(all), 'triggers on pull_request');\nconsole.log('CIPIPE_ORACLE triggers ok');\n\n// No invented secrets, and no publish job in a repo with no release evidence.\nfor (const m of all.matchAll(/secrets\\.([A-Za-z0-9_]+)/g)) {\n  assert.equal(m[1], 'GITHUB_TOKEN', `workflow references invented secret \"${m[1]}\"`);\n}\nassert.ok(!/npm\\s+publish/.test(all), 'no publish step - the repo shows no release configuration');\nconsole.log('CIPIPE_ORACLE nothing invented');\nconsole.log('CIPIPE_ORACLE done');\n",
+        "surface": "harness"
+      }
+    ]
+  },
+  "mocks": [],
+  "success": {
+    "summary": "A real GitHub Actions workflow exists at the conventional location, runs exactly the repo's two scripts on push and pull_request with pinned actions, invents no command, secret, or publish job — and every fleet artifact exists with the workflow commands verified against quoted package.json lines.",
+    "deliverables": [
+      {
+        "path": "{{task.dir}}/ci-plan.md",
+        "kind": "markdown-notes",
+        "artifact": true,
+        "minBytes": 500,
+        "checks": [
+          {
+            "kind": "contains",
+            "file": "{{task.dir}}/ci-plan.md",
+            "pattern": "^##\\s+Detected commands[\\s\\S]*^##\\s+Pipeline stages[\\s\\S]*^##\\s+Caching[\\s\\S]*^##\\s+Triggers",
+            "flags": "im",
+            "label": "ci-plan sections in order"
+          },
+          {
+            "kind": "contains",
+            "file": "{{task.dir}}/ci-plan.md",
+            "pattern": "package\\.json",
+            "label": "the plan reads the real manifest"
+          },
+          {
+            "kind": "citationsResolve",
+            "file": "{{task.dir}}/ci-plan.md",
+            "minCitations": 2,
+            "artifact": true
+          }
+        ]
+      },
+      {
+        "path": "{{task.dir}}/ci-notes.md",
+        "kind": "markdown-notes",
+        "artifact": true,
+        "minBytes": 500,
+        "checks": [
+          {
+            "kind": "contains",
+            "file": "{{task.dir}}/ci-notes.md",
+            "pattern": "^##\\s+Files written[\\s\\S]*^##\\s+Stages[\\s\\S]*^##\\s+Choices explained",
+            "flags": "im",
+            "label": "ci-notes sections in order"
+          },
+          {
+            "kind": "contains",
+            "file": "{{task.dir}}/ci-notes.md",
+            "pattern": "\\.github/workflows/",
+            "label": "names the conventional workflow location"
+          },
+          {
+            "kind": "citationsResolve",
+            "file": "{{task.dir}}/ci-notes.md",
+            "minCitations": 2,
+            "artifact": true
+          }
+        ]
+      },
+      {
+        "path": "{{task.dir}}/verification.md",
+        "kind": "markdown-notes",
+        "artifact": true,
+        "minBytes": 500,
+        "checks": [
+          {
+            "kind": "contains",
+            "file": "{{task.dir}}/verification.md",
+            "pattern": "^##\\s+Workflow files[\\s\\S]*^##\\s+Commands verified against package\\.json[\\s\\S]*^##\\s+Result",
+            "flags": "im",
+            "label": "verification sections in order"
+          }
+        ]
+      },
+      {
+        "path": "{{task.dir}}/review.md",
+        "kind": "markdown-report",
+        "artifact": true,
+        "minBytes": 400,
+        "checks": [
+          {
+            "kind": "contains",
+            "file": "{{task.dir}}/review.md",
+            "pattern": "Verdict:\\s*(?:PASS|REVISE)",
+            "flags": "i",
+            "label": "explicit reviewer verdict"
+          }
+        ]
+      }
+    ],
+    "checks": [
+      {
+        "kind": "nodeRuns",
+        "file": "tests/greet.test.mjs"
+      },
+      {
+        "kind": "nodeScriptPasses",
+        "script": "tools/lint.mjs",
+        "timeoutMs": 30000,
+        "requiredOutput": [
+          {
+            "pattern": "lint ok",
+            "label": "the repo lint still passes"
+          }
+        ]
+      },
+      {
+        "kind": "nodeScriptPasses",
+        "script": "tests/verify-ci.mjs",
+        "timeoutMs": 60000,
+        "requiredOutput": [
+          {
+            "pattern": "CIPIPE_ORACLE workflow found",
+            "label": "a workflow exists at the conventional location"
+          },
+          {
+            "pattern": "CIPIPE_ORACLE scripts ok",
+            "label": "both real scripts run in CI"
+          },
+          {
+            "pattern": "CIPIPE_ORACLE no invented commands",
+            "label": "no command invented beyond package.json"
+          },
+          {
+            "pattern": "CIPIPE_ORACLE triggers ok",
+            "label": "triggers on push and pull_request"
+          },
+          {
+            "pattern": "CIPIPE_ORACLE nothing invented",
+            "label": "no fabricated secrets or publish job"
+          }
+        ]
+      }
+    ],
+    "taskNotes": {
+      "minBytes": 120,
+      "checks": [
+        {
+          "kind": "contains",
+          "file": "task-notes.md",
+          "pattern": "\\bDONE\\b[\\s\\S]*(workflows|ci\\.yml|verification\\.md)",
+          "flags": "i",
+          "label": "DONE note names the workflow deliverable"
+        }
+      ]
+    },
+    "taskGraph": {
+      "requireCraftbookTask": true,
+      "requireTerminalStep": true
+    },
+    "unchangedFixtures": [
+      "package.json",
+      "package-lock.json",
+      "src/greet.js",
+      "tools/lint.mjs",
+      "tests/greet.test.mjs"
+    ]
+  },
+  "rubric": {
+    "artifact": {
+      "path": "{{task.dir}}/ci-notes.md",
+      "kind": "markdown"
+    },
+    "axes": [
+      {
+        "name": "Grounded inventory",
+        "description": "Every planned stage maps to a script that exists in package.json, quoted verbatim; no typecheck, build, or publish stage was invented for this repo."
+      },
+      {
+        "name": "Workflow quality",
+        "description": "The workflow sits at the conventional location, pins actions to major versions, orders install before lint and test, wires dependency caching, and triggers on push and pull_request."
+      },
+      {
+        "name": "Verification honesty",
+        "description": "The verify pass matched each workflow command to a quoted manifest line and is plain about what only a live CI run on the host can prove."
+      }
+    ]
+  },
+  "qualityFocus": [
+    "no invented commands or secrets",
+    "manifest-quoted verification",
+    "enforced review loop"
+  ]
+}

package/data/craftbook-templates/ci/citation-audit/versions/1.0.3/craftbook.json

@@ -0,0 +1,278 @@
+{
+  "id": "citation-audit",
+  "name": "Citation Audit",
+  "description": "Audit an existing document's claims against its citations and report every mismatch: a reviewer pairs each cited claim with the citation it depends on and locks the audit rubric, a researcher opens each cited source and verifies the claim is actually supported, and a reviewer compiles a per-citation audit report with verdicts and fixes. The ordering matters because auditing without first pairing claims to citations misses unsupported claims, and a report without a fixed verdict scale is inconsistent; building the claim-citation pairing and rubric first makes the audit exhaustive and the report actionable. Use this for checking that citations support their claims, source-verification of a draft, catching citation drift or unsupported statements, and 'do these references actually back up the text' reviews.\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 audit (reviewer) — pair each cited claim to its citation + lock the rubric → gated on artifact `{{workPath}}/scope.md` (markdown-notes)\n2. Verify each citation (researcher) — open each source, confirm it backs the paired claim → gated on artifact `{{workPath}}/verify.md` (markdown-notes)\n3. Compile the audit report (reviewer) — per-citation verdicts, justifying quotes, and fixes → gated on artifact `{{workPath}}/audit.md` (markdown-report)\n\nSource discipline:\n\n- Audit only claims that appear in the supplied draft or source bundle.\n- Keep the draft's citation labels verbatim, such as `Claim [A]`, so each verdict is traceable.\n- Treat uncited factual sentences in the draft as `Uncited-Claim`; do not create new claims to make the report longer.\n- If a source excerpt is missing or contradicts the draft, say that directly and propose a correction or citation fix.\n\nThe gates never advance with an unmet criterion, and loop back to the\nowning phase to fix named gaps.\n",
+  "entryStepId": "scope",
+  "triggers": [
+    "audit the citations",
+    "check citations support the claims",
+    "verify the references in this doc",
+    "citation accuracy check",
+    "are these sources legit"
+  ],
+  "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 audit",
+      "description": "pair each cited claim to its citation + lock the rubric",
+      "prompt": "1. Read the target document and build a pairing table in {{workPath}}/scope.md: for every inline citation, record a pair id (P1, P2, ...), the exact claim sentence it supports, and the citation/reference it points to. 2. Also flag every factual claim that has NO citation as a candidate (these are audit findings too). 3. Lock the verdict rubric: Supported (source clearly backs the claim), Partially-Supported (source backs part / overstates), Not-Supported (source does not back it), Source-Missing/Dead (citation cannot be located), Uncited-Claim. 4. Write an acceptance-criteria checklist (e.g. 'every inline citation is paired and audited', 'uncited factual claims are listed', 'each pair gets exactly one verdict from the rubric', 'each Not-Supported/Partial verdict quotes the source to justify it', 'a fix is proposed per problem'). 5. Record decisions. Call write_task_note with the pairing table + rubric + checklist; write the same to {{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": "reviewer",
+      "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": "verify",
+      "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": "verify",
+      "name": "Verify each citation",
+      "description": "open each source, confirm it backs the paired claim",
+      "prompt": "1. For each pair id, locate and open the cited source and read the relevant part. 2. Log to {{workPath}}/verify.md: the pair id, whether the source was reachable, the exact passage/figure the claim relies on (quote it), and a proposed verdict from the locked rubric with a one-line reason. 3. Be strict: a source on the right topic that does not actually state the specific claim is Not-Supported, not Supported. 4. For dead or unfindable citations, record Source-Missing/Dead and what you searched. 5. For each problem, draft a concrete fix (correct the claim, swap the source, or add a citation). Good looks like a verify log where every pair's verdict is justified by a quote. Do not write the final report yet.\n\nThe deliverable `{{workPath}}/verify.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": "researcher",
+      "advanceWhen": {
+        "file": "{{workPath}}/verify.md",
+        "minBytes": 1,
+        "sniff": "nonempty",
+        "artifact": true
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "{{workPath}}/verify.md",
+            "bytes": 120,
+            "artifact": true
+          },
+          {
+            "kind": "sniff",
+            "file": "{{workPath}}/verify.md",
+            "sniff": "nonempty",
+            "artifact": true
+          }
+        ],
+        "onReject": "verify",
+        "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",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "artifact"
+      }
+    },
+    {
+      "id": "report",
+      "name": "Compile the audit report",
+      "description": "per-citation verdicts, justifying quotes, and fixes",
+      "prompt": "1. Write {{workPath}}/audit.md with a summary (counts per verdict, and the count of uncited claims) then one entry per pair id: the claim, the citation, the verdict from the rubric, the justifying quote from the source, and the proposed fix for any non-Supported verdict. 2. Add a dedicated 'Uncited claims' section listing factual statements that lack a citation. 3. Use ONLY rubric verdicts and base each strictly on {{workPath}}/verify.md. 4. Make it actionable — an editor should be able to apply every fix directly. 5. On a loop-back, fix only the named gaps. 6. Call write_task_note with the report path and which criteria now pass.\n\nBefore working, open the artifact input `{{workPath}}/verify.md` with `read_artifact`; do not look for it in the workspace.\n\nThe deliverable `{{workPath}}/audit.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": "reviewer",
+      "consumes": [
+        {
+          "file": "{{workPath}}/verify.md",
+          "artifact": true
+        }
+      ],
+      "advanceWhen": {
+        "file": "{{workPath}}/audit.md",
+        "minBytes": 1,
+        "sniff": "nonempty",
+        "artifact": true
+      },
+      "gate": {
+        "at": "completion",
+        "checks": [
+          {
+            "kind": "minBytes",
+            "file": "{{workPath}}/audit.md",
+            "bytes": 1500,
+            "artifact": true
+          },
+          {
+            "kind": "contains",
+            "file": "{{workPath}}/audit.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",
+          "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": "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}}/audit.md and check EACH criterion PASS/FAIL: (1) every citation paired in {{workPath}}/scope.md has an audited entry — none skipped; (2) an uncited-claims section lists statements lacking citations; (3) each entry has exactly one rubric verdict; (4) every Not-Supported/Partially-Supported/Dead verdict is justified by a quoted source passage (spot-check 3 against {{workPath}}/verify.md); (5) each problem has a concrete, applicable fix; (6) the summary counts match the entries. A report that marks a topically-related-but-non-supporting source as Supported, or skips a citation, FAILS.\n\nThe deliverable lives in the project's artifacts drawer — open `{{workPath}}/audit.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: \"report\" })` 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}}/audit.md",
+          "artifact": true
+        }
+      ],
+      "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": "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/ci/citation-audit/versions/1.0.3/test.json

@@ -0,0 +1,152 @@
+{
+  "schemaVersion": 1,
+  "title": "Citation audit over seeded draft and source excerpts",
+  "objective": "Measure whether the citation-audit craftbook pairs claims to citations and reports supported, partial, unsupported, dead, and uncited claims.",
+  "tags": [
+    "corpus"
+  ],
+  "prompt": "In the `Citation Audit Eval` project, read only `source/draft.md` and `source/source-excerpts.md`, then write `tasks/eval/audit.md`. Audit exactly the four cited claims `Claim [A]` through `Claim [D]`, plus the uncited sentence about Finance receiving a weekly exception export. Do not add Claim [E], Claim [F], outside company metrics, or imagined source sections. Use verdicts Supported, Not-Supported, Source-Missing/Dead, and Uncited-Claim, with source quotes and concrete fixes.",
+  "setup": {
+    "projectName": "Citation Audit Eval",
+    "about": "Self-contained eval project for citation-audit. Draft and source excerpts are seeded under workspace/source.",
+    "missionObjectives": "Audit every citation and uncited factual claim with strict verdicts, quotes from source excerpts, and concrete fixes.",
+    "files": [
+      {
+        "path": "source/draft.md",
+        "content": "# Draft\n\nGuided returns intake reduced median first response from 18 hours to 6 hours [A]. The pilot covered every SKU in the catalog [B]. Customers were 34% more satisfied after photo evidence was required [C]. Finance now receives a weekly exception export. Warehouse teams need barcode exception training before national rollout [D].\n"
+      },
+      {
+        "path": "source/source-excerpts.md",
+        "content": "# Source excerpts\n\n[A] Pilot dashboard: median first response improved from 18 hours to 6 hours for the 18 pilot SKUs.\n[B] Rollout memo: March pilot covered 18 home-office accessory SKUs, not the full catalog.\n[C] Source missing: no satisfaction survey was run during the pilot.\n[D] Operations notes: warehouse teams need barcode exception training before national rollout.\n"
+      }
+    ],
+    "worker": {
+      "name": "Iris",
+      "role": "Reviewer"
+    },
+    "craftbookParams": {
+      "workPath": "tasks/eval"
+    }
+  },
+  "mocks": [],
+  "success": {
+    "summary": "workspace/audit.md audits all seeded citations and uncited claims with rubric verdicts, source quotes, counts, and fixes.",
+    "deliverables": [
+      {
+        "path": "tasks/eval/audit.md",
+        "kind": "markdown-report",
+        "minBytes": 1200,
+        "checks": [
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "(?:^|\\n)#{1,3}\\s+\\S",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "Supported|Partially-Supported|Not-Supported|Source-Missing|Uncited-Claim",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "\\[A\\]|\\[B\\]|\\[C\\]|\\[D\\]",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "Claim \\[A\\][\\s\\S]{0,500}Supported",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "Claim \\[B\\][\\s\\S]{0,500}Not-Supported",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "Claim \\[C\\][\\s\\S]{0,500}Source-Missing|Claim \\[C\\][\\s\\S]{0,500}Dead",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "Claim \\[D\\][\\s\\S]{0,500}Supported",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "Finance[\\s\\S]{0,500}Uncited-Claim|Uncited-Claim[\\s\\S]{0,500}Finance",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "18\\s*hours|6\\s*hours|18 home-office accessory SKUs|full catalog",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "34%|satisfaction survey",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "Finance|weekly exception export",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "fix|correction|replace|add a citation",
+            "flags": "i"
+          },
+          {
+            "kind": "notContains",
+            "file": "tasks/eval/audit.md",
+            "pattern": "Claim \\[E\\]|Claim \\[F\\]|cost per return|50 concurrent|RESTful API|Q3|stress tests|Financial summary|System architecture|expansion of this report",
+            "flags": "i",
+            "label": "do not invent audit claims or source excerpts beyond the seeded draft"
+          }
+        ]
+      }
+    ]
+  },
+  "rubric": {
+    "artifact": {
+      "path": "tasks/eval/audit.md",
+      "kind": "markdown"
+    },
+    "axes": [
+      {
+        "name": "grounding",
+        "description": "Seeded facts appear faithfully and the source is cited."
+      },
+      {
+        "name": "structure",
+        "description": "Sections progress logically for the artifact class."
+      },
+      {
+        "name": "completeness",
+        "description": "Required elements (risks, next actions, scope) are all present."
+      },
+      {
+        "name": "tone",
+        "description": "The register fits the audience and artifact class."
+      }
+    ]
+  },
+  "qualityFocus": [
+    "claim-citation pairing",
+    "strict verdicts",
+    "actionable fixes"
+  ]
+}

package/data/craftbook-templates/cl/cli-tool/versions/1.1.2/test.json

@@ -14,9 +14,6 @@
     "worker": {
       "name": "Koen",
       "role": "Developer"
-    },
-    "craftbookParams": {
-      "workPath": "tasks/eval"
     }
   },
   "mocks": [],