@bendyline/gilde 0.1.50 → 0.1.52

package/data/craftbook-templates/sh/ship/versions/1.1.1/craftbook.json

@@ -0,0 +1,301 @@
+{
+  "id": "ship",
+  "name": "Ship",
+  "description": "End-to-end release procedure. Six normal\nsteps + one halt step the branches divert to on failure:\n\n1. **Preflight** — snapshot branch + working-tree state, gate trunk pushes\n2. **Run tests** — project's primary test command\n3. **Self-review pass** — last walk through the diff\n4. **Open PR** — push + create\n5. **Wait for CI** — poll until green or failure\n6. **Summary** — final DONE\n7. **Halt** (branched target) — DONE_WITH_CONCERNS on any failure\n\nThe branches are the load-bearing piece. If tests fail, CI fails, or\nthe poll times out, the recipe routes to `halt` rather than blasting\nthrough. The upstream procedure encoded that safety property in\nimperative if/else; the craftbook schema makes it explicit.\n\nScripts:\n- `run-tests.ts` — wraps `run_package_script` with the right test command\n- `open-pr.ts` — pushes the branch, calls `github_pr_create`\n- `wait-checks.ts` — polls `github_check_status` until settled\n\n## Applicability\n\nThe craftbook declares two manifest-level `requirements` so the launcher\nand command registry only offer it where it can actually run:\n\n- **github** — the project must be connected to a GitHub repo (the\n  `open-pr` and `wait-checks` scripts call `github_pr_create` /\n  `github_check_status` directly; without a connected repo they have\n  nowhere to push to).\n- **non-main-branch** — the project's checkout must be on a branch\n  other than `main` / `master`. The preflight step's `ask_user_question`\n  was a runtime guard against accidental trunk shipping; promoting it\n  to a hard `requirements` gate removes the temptation entirely. The\n  step-level prompt still re-checks in case the catalog's branch\n  snapshot is stale.\n\nRequires the github toolset configured. Both gates evaluate via\n`unmetCraftbookRequirements` in `@bendyline/gezel` and are surfaced\nthrough `/api/projects/:id/craftbooks` so the Commands panel hides the\ncraftbook in projects that don't qualify rather than letting the user\nlaunch it and watch step 4 fail.\n",
+  "basedOn": {
+    "name": "gstack",
+    "url": "https://github.com/garrytan/gstack"
+  },
+  "entryStepId": "preflight",
+  "triggers": [
+    "ship this",
+    "ship it",
+    "open a pr",
+    "open the pr",
+    "release this"
+  ],
+  "requirements": [
+    {
+      "kind": "github"
+    },
+    {
+      "kind": "non-main-branch"
+    }
+  ],
+  "steps": [
+    {
+      "id": "preflight",
+      "name": "Preflight",
+      "description": "Confirm the working tree is in a shippable state: branch is right, no uncommitted weirdness, what is/isn't included.",
+      "prompt": "Before anything destructive, snapshot the state:\n\n1. Call `inspect_git_workdir` (or equivalent) for the current branch, dirty status, ahead/behind.\n2. Read the most recent commit messages on this branch to confirm scope.\n3. If the branch is `main` or the configured trunk, STOP and use `ask_user_question` — shipping straight to main needs explicit confirmation.\n4. If the working tree is dirty, ask whether to commit first or stash.\n\nWrite the snapshot to task notes. Advance.",
+      "suggestedRole": "developer",
+      "next": "run-tests",
+      "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": "run-tests",
+      "name": "Run tests",
+      "description": "Run the project's test command. Branches on failure to a triage step; on success advances to the review pass.",
+      "prompt": "Run the `run-tests` script. It picks the project's primary test command (npm `test` script, `pnpm test`, etc.) and runs it. Output: `{ passed: boolean, summary: string }`.\n\nIf passed: advance.\nIf failed: write the failure summary to task notes and DO NOT proceed. Report DONE_WITH_CONCERNS with the failure summary — let the user fix and re-run.",
+      "suggestedRole": "developer",
+      "onExit": {
+        "name": "run-tests",
+        "autoAdvanceWhen": {
+          "op": "equals",
+          "field": "passed",
+          "value": true
+        }
+      },
+      "next": "review-pass",
+      "branches": [
+        {
+          "when": {
+            "op": "equals",
+            "field": "passed",
+            "value": false
+          },
+          "goto": "halt"
+        }
+      ],
+      "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",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "task-note"
+      }
+    },
+    {
+      "id": "review-pass",
+      "name": "Self-review pass",
+      "description": "Walk your own diff before opening the PR. Last chance to catch obvious mistakes.",
+      "prompt": "Read your own diff (`git diff @{u}...` or against the base branch). Look for:\n\n- Debug prints / console.log left in\n- Commented-out code\n- Filenames that don't match the rename you intended\n- Test files you forgot to commit\n- Migration files in the wrong direction\n\nIf anything jumps out, fix it now via `write_file` + commit. If clean, advance.",
+      "suggestedRole": "reviewer",
+      "next": "open-pr",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web"
+        ],
+        "outputMedium": "workspace"
+      }
+    },
+    {
+      "id": "open-pr",
+      "name": "Open PR",
+      "description": "Push the branch + open the pull request.",
+      "prompt": "Run the `open-pr` script. It pushes the current branch (via `pushProjectGithub`-equivalent) and calls `github_pr_create` with a title derived from the most recent commit and a body that captures the scope.\n\nIf the script returns an error (push rejected, branch already has an open PR, missing PAT), surface the message and stop here — don't try to guess past it.",
+      "suggestedRole": "developer",
+      "onExit": {
+        "name": "open-pr",
+        "autoAdvanceWhen": {
+          "op": "exists",
+          "field": "url"
+        }
+      },
+      "next": "wait-checks",
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "none"
+      }
+    },
+    {
+      "id": "wait-checks",
+      "name": "Wait for CI",
+      "description": "Poll workflow runs / check status until the gate is green or fails.",
+      "prompt": "Run the `wait-checks` script with the branch name and a sensible timeout (default 10 minutes). It polls `github_workflow_runs` + `github_check_status` until the state settles to success / failure / timeout.\n\nWhile polling, do not chat — wait. When it returns:\n\n- `success`: advance to summary.\n- `failure`: write the failing checks to task notes; report DONE_WITH_CONCERNS.\n- `timeout`: report status, let the user decide whether to keep waiting.",
+      "suggestedRole": "developer",
+      "onExit": {
+        "name": "wait-checks",
+        "autoAdvanceWhen": {
+          "op": "equals",
+          "field": "state",
+          "value": "success"
+        }
+      },
+      "next": "summary",
+      "branches": [
+        {
+          "when": {
+            "op": "equals",
+            "field": "state",
+            "value": "failure"
+          },
+          "goto": "halt"
+        },
+        {
+          "when": {
+            "op": "equals",
+            "field": "state",
+            "value": "timeout"
+          },
+          "goto": "halt"
+        }
+      ],
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "task-note"
+      }
+    },
+    {
+      "id": "summary",
+      "name": "Summary",
+      "description": "Stamp a final summary. PR opened, CI green, ready to merge.",
+      "prompt": "Write a one-paragraph DONE summary: PR number, URL, scope (one line), test result, CI verdict. Save to task notes. Report DONE.",
+      "suggestedRole": "developer",
+      "terminal": true,
+      "toolPolicy": {
+        "disallowBuiltinToolsets": [
+          "ai-apps",
+          "archives",
+          "artifacts",
+          "audio",
+          "browser-automation",
+          "code-execution",
+          "craftbooks",
+          "data-tables",
+          "entity-intel",
+          "image-intel",
+          "images",
+          "role-delegation",
+          "role-delegation-escalation",
+          "security-intel",
+          "team-management",
+          "videos",
+          "web",
+          "workspace-fs-write"
+        ],
+        "outputMedium": "none"
+      }
+    },
+    {
+      "id": "halt",
+      "name": "Halted",
+      "description": "Something blocked the ship — tests, CI, or push. Report DONE_WITH_CONCERNS.",
+      "prompt": "Report DONE_WITH_CONCERNS with the specific failure surfaced upstream. Do not retry, do not push past — the user needs to see this and decide.",
+      "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": "none"
+      }
+    }
+  ],
+  "scripts": {
+    "run-tests": "import { defineScript, gezel } from '@bendyline/gezel-sdk';\n\nexport const meta = defineScript({\n  name: 'run-tests',\n  description:\n    'Run the project test command via run_package_script. Stamps {passed: boolean, summary: string}.',\n  inputs: {\n    scriptName: {\n      type: 'string',\n      description: 'Override the npm script name. Defaults to \"test\".',\n      default: 'test',\n    },\n  },\n  outputs: {\n    passed: { type: 'boolean', description: 'True iff the test command exited 0.' },\n    summary: { type: 'string', description: 'Last ~2KB of stdout/stderr.' },\n  },\n  requires: ['workspace.read'],\n});\n\ninterface PackageScriptResult {\n  exitCode: number;\n  stdout?: string;\n  stderr?: string;\n}\n\nasync function main(): Promise<void> {\n  const input = gezel.input as { scriptName?: string };\n  const scriptName = input.scriptName ?? 'test';\n  const raw = await gezel.mcp.call('run_package_script', { script: scriptName });\n  let parsed: PackageScriptResult | null = null;\n  if (typeof raw === 'string') {\n    try {\n      parsed = JSON.parse(raw) as PackageScriptResult;\n    } catch {\n      // The MCP tool returns a text blob, not always JSON. We treat\n      // \"non-zero exit code\" / \"failed\" / \"FAIL \" markers as failure.\n    }\n  }\n  const text = typeof raw === 'string' ? raw : JSON.stringify(raw);\n  const looksFailed =\n    parsed?.exitCode !== undefined\n      ? parsed.exitCode !== 0\n      : /(?:^|\\b)(?:fail|error|exit code [1-9]|✗)\\b/i.test(text);\n  const tail = text.length > 2048 ? text.slice(-2048) : text;\n  gezel.log(`pnpm ${scriptName}: ${looksFailed ? 'failed' : 'passed'}`);\n  gezel.output({ passed: !looksFailed, summary: tail });\n}\n\nawait main();\n",
+    "open-pr": "import { defineScript, gezel } from '@bendyline/gezel-sdk';\n\nexport const meta = defineScript({\n  name: 'open-pr',\n  description:\n    'Open a pull request for the current branch via the GitHub MCP tools. Stamps {number, url} on success.',\n  inputs: {\n    title: { type: 'string', description: 'PR title. Falls back to a derived placeholder.' },\n    body: { type: 'string', description: 'PR body. Markdown.' },\n    head: {\n      type: 'string',\n      description: 'Source branch. Defaults to the current branch (caller must resolve).',\n      required: true,\n    },\n    base: { type: 'string', description: 'Target branch.', default: 'main' },\n    draft: { type: 'boolean', description: 'Open as draft.', default: false },\n  },\n  outputs: {\n    number: { type: 'number', description: 'Created PR number.' },\n    url: { type: 'string', description: 'PR URL.' },\n  },\n  requires: ['network'],\n});\n\nasync function main(): Promise<void> {\n  const input = gezel.input as {\n    title?: string;\n    body?: string;\n    head: string;\n    base?: string;\n    draft?: boolean;\n  };\n  const title = input.title ?? `Ship: ${input.head}`;\n  const args: Record<string, unknown> = {\n    title,\n    head: input.head,\n    base: input.base ?? 'main',\n  };\n  if (input.body) args.body = input.body;\n  if (input.draft !== undefined) args.draft = input.draft;\n  const raw = await gezel.mcp.call('github_pr_create', args);\n  const text = typeof raw === 'string' ? raw : '';\n  const m = /Opened PR #(\\d+) — (\\S+)/.exec(text);\n  if (!m) {\n    throw new Error(`unexpected github_pr_create response: ${text.slice(0, 200)}`);\n  }\n  gezel.output({ number: Number(m[1]), url: m[2]! });\n}\n\nawait main();\n",
+    "wait-checks": "import { defineScript, gezel } from '@bendyline/gezel-sdk';\n\nexport const meta = defineScript({\n  name: 'wait-checks',\n  description:\n    'Poll github_check_status for a ref until it settles. Returns {state: success|failure|pending|timeout, checks[]}.',\n  inputs: {\n    ref: { type: 'string', description: 'Branch name or commit sha.', required: true },\n    timeoutMs: {\n      type: 'number',\n      description: 'How long to wait before giving up.',\n      default: 600_000,\n      integer: true,\n    },\n    pollIntervalMs: {\n      type: 'number',\n      description: 'How long between polls.',\n      default: 15_000,\n      integer: true,\n    },\n  },\n  outputs: {\n    state: { type: 'string', description: 'Final state: success / failure / timeout.' },\n    checks: { type: 'array', description: 'Per-check breakdown.', itemType: 'object' },\n  },\n  requires: ['network'],\n});\n\ninterface CheckBlock {\n  name: string;\n  status: string;\n  conclusion: string | null;\n  url?: string;\n}\n\nfunction parseChecks(text: string): { state: string; checks: CheckBlock[] } {\n  const lines = text.split('\\n');\n  const stateLine = lines[0] ?? '';\n  const stateMatch = /^State:\\s*(\\w+)/i.exec(stateLine);\n  const state = stateMatch?.[1]?.toLowerCase() ?? 'unknown';\n  const checks: CheckBlock[] = [];\n  for (let i = 1; i < lines.length; i++) {\n    const m = /^• (.+?) — (\\S+?)(?:\\s*\\/\\s*(\\S+))?$/.exec(lines[i] ?? '');\n    if (!m) continue;\n    checks.push({\n      name: m[1]!,\n      status: m[2]!,\n      conclusion: m[3] ?? null,\n    });\n  }\n  return { state, checks };\n}\n\nasync function main(): Promise<void> {\n  const input = gezel.input as { ref: string; timeoutMs?: number; pollIntervalMs?: number };\n  const deadline = Date.now() + (input.timeoutMs ?? 600_000);\n  const poll = input.pollIntervalMs ?? 15_000;\n  let lastState = 'unknown';\n  let lastChecks: CheckBlock[] = [];\n  while (Date.now() < deadline) {\n    const raw = await gezel.mcp.call('github_check_status', { ref: input.ref });\n    const text = typeof raw === 'string' ? raw : '';\n    const parsed = parseChecks(text);\n    lastState = parsed.state;\n    lastChecks = parsed.checks;\n    gezel.log(`checks for ${input.ref}: ${parsed.state} (${parsed.checks.length} check(s))`);\n    if (parsed.state === 'success' || parsed.state === 'failure') {\n      gezel.output({ state: parsed.state, checks: parsed.checks });\n      return;\n    }\n    await new Promise((r) => setTimeout(r, poll));\n  }\n  gezel.output({ state: 'timeout', checks: lastChecks });\n}\n\nawait main();\n"
+  },
+  "version": "1.1.1",
+  "releasedAt": "2026-09-01T20:00:00Z"
+}

package/data/craftbook-templates/sh/ship/versions/1.1.1/test.json

@@ -0,0 +1,225 @@
+{
+  "schemaVersion": 1,
+  "title": "Ship flow against fake GitHub/CI",
+  "objective": "Exercise the ship craftbook without touching a real remote by replacing git/GitHub/CI with a deterministic test CLI and MCP surface.",
+  "tags": [
+    "external"
+  ],
+  "prompt": "Time to ship the packing-count export. Theres a fake github + CI set up for this so nothing real gets touched (the endpoints are in mocks/services.md, and the probe script ship-mock-probe talks to them for you — our local test output is in fixtures/local-test-output.log). Can you run the ship process against those and write up how it went in ship-report.md, plus the PR details in fake-pr.json? Branch is chore/packing-count-export, should end up as PR 42.",
+  "setup": {
+    "projectName": "Ship Eval",
+    "about": "Self-contained eval project for ship craftbook behavior. Fixtures emulate GitHub, local tests, and CI.",
+    "missionObjectives": "Use only the fake fixtures. Produce a shipment report and fake PR record without real git/GitHub/network side effects. Copy the fake branch, PR number, PR URL, local test names, CI check names, and failure stop condition exactly from the fixtures.",
+    "files": [
+      {
+        "path": "fixtures/local-test-output.log",
+        "content": "$ npm test\nlint: pass\nunit: pass (42 tests)\nplaywright-smoke: pass\nduration: 18.4s"
+      }
+    ],
+    "worker": {
+      "name": "Ruben",
+      "role": "Developer"
+    }
+  },
+  "mocks": [
+    {
+      "kind": "http",
+      "id": "github",
+      "description": "Fake GitHub API for the co-op/returns-desk repository. POST a pull request for branch chore/packing-count-export; GET it back by number.",
+      "credential": {
+        "name": "mock.github"
+      },
+      "routes": [
+        {
+          "method": "POST",
+          "path": "/repos/co-op/returns-desk/pulls",
+          "status": 201,
+          "body": {
+            "number": 42,
+            "state": "open",
+            "branch": "chore/packing-count-export",
+            "html_url": "https://fake.github.local/co-op/returns-desk/pull/42"
+          }
+        },
+        {
+          "method": "GET",
+          "path": "/repos/co-op/returns-desk/pulls/42",
+          "body": {
+            "number": 42,
+            "state": "open",
+            "branch": "chore/packing-count-export",
+            "html_url": "https://fake.github.local/co-op/returns-desk/pull/42"
+          }
+        }
+      ]
+    },
+    {
+      "kind": "http",
+      "id": "ci",
+      "description": "Fake CI API. Poll GET /runs/pr-42 for the pipeline attached to the fake PR.",
+      "credential": {
+        "name": "mock.ci"
+      },
+      "routes": [
+        {
+          "method": "GET",
+          "path": "/runs/pr-42",
+          "body": {
+            "run": "pr-42",
+            "status": "success",
+            "checks": [
+              {
+                "name": "lint",
+                "status": "success"
+              },
+              {
+                "name": "unit",
+                "status": "success"
+              },
+              {
+                "name": "playwright-smoke",
+                "status": "success"
+              }
+            ]
+          }
+        },
+        {
+          "method": "GET",
+          "path": "/runs/*",
+          "body": {
+            "status": "pending"
+          }
+        }
+      ]
+    },
+    {
+      "kind": "cli",
+      "id": "probe",
+      "description": "Provenance-trusted project script `ship-mock-probe` that exercises both fake services through http.authed and returns their raw responses.",
+      "shim": {
+        "path": "scripts/ship-mock-probe.ts",
+        "content": "import { defineScript, gezel } from '@bendyline/gezel-sdk';\n\nexport const meta = defineScript({\n  name: 'ship-mock-probe',\n  description:\n    'Probe the live fake GitHub/CI services for this eval: opens the fake PR for branch chore/packing-count-export and polls CI for run pr-42. Outputs the raw JSON responses for the ship report.',\n  inputs: {},\n  outputs: {\n    pr: { type: 'string', description: 'Raw JSON body from the fake GitHub PR creation.' },\n    ci: { type: 'string', description: 'Raw JSON body from the fake CI status poll.' },\n  },\n  requires: ['workspace.read', 'network', 'credential:mock.github', 'credential:mock.ci'],\n});\n\nconst services = JSON.parse(await gezel.fs.read('mocks/services.json'));\nconst base = (id) => {\n  const service = services.find((entry) => entry.id === id);\n  if (!service) throw new Error(`mock service \"${id}\" is not listed in mocks/services.json`);\n  return service.baseUrl;\n};\n\nconst pr = await gezel.http.authed(`${base('github')}/repos/co-op/returns-desk/pulls`, {\n  credential: 'mock.github',\n  method: 'POST',\n  body: JSON.stringify({ branch: 'chore/packing-count-export', title: 'Packing count export' }),\n});\nif (!pr.ok) throw new Error(`fake GitHub PR creation failed: ${pr.status} ${pr.body}`);\n\nconst ci = await gezel.http.authed(`${base('ci')}/runs/pr-42`, { credential: 'mock.ci' });\nif (!ci.ok) throw new Error(`fake CI poll failed: ${ci.status} ${ci.body}`);\n\ngezel.output({ pr: pr.body, ci: ci.body });\n"
+      }
+    }
+  ],
+  "success": {
+    "summary": "ship-report.md + fake-pr.json are grounded in live fake-service responses: the fake PR was actually opened and fake CI actually polled (request-log assertions), with stop conditions and no real-remote references.",
+    "deliverables": [
+      {
+        "path": "ship-report.md",
+        "kind": "markdown-report",
+        "minBytes": 1200,
+        "checks": [
+          {
+            "kind": "contains",
+            "file": "ship-report.md",
+            "pattern": "preflight|local tests|npm test",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "ship-report.md",
+            "pattern": "lint|unit|playwright-smoke",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "ship-report.md",
+            "pattern": "https://fake\\.github\\.local/co-op/returns-desk/pull/42"
+          },
+          {
+            "kind": "contains",
+            "file": "ship-report.md",
+            "pattern": "chore/packing-count-export"
+          },
+          {
+            "kind": "contains",
+            "file": "ship-report.md",
+            "pattern": "CI|checks|green|success",
+            "flags": "i"
+          },
+          {
+            "kind": "contains",
+            "file": "ship-report.md",
+            "pattern": "stop|failure|do not ship",
+            "flags": "i"
+          },
+          {
+            "kind": "notContains",
+            "file": "ship-report.md",
+            "pattern": "github\\.com|real remote|pushed to origin",
+            "flags": "i"
+          }
+        ]
+      },
+      {
+        "path": "fake-pr.json",
+        "kind": "json",
+        "minBytes": 180,
+        "checks": [
+          {
+            "kind": "contains",
+            "file": "fake-pr.json",
+            "pattern": "42"
+          },
+          {
+            "kind": "contains",
+            "file": "fake-pr.json",
+            "pattern": "chore/packing-count-export"
+          },
+          {
+            "kind": "contains",
+            "file": "fake-pr.json",
+            "pattern": "https://fake\\.github\\.local/co-op/returns-desk/pull/42"
+          }
+        ]
+      }
+    ],
+    "mocks": [
+      {
+        "service": "github",
+        "minRequests": 1,
+        "requiredPaths": [
+          "^/repos/co-op/returns-desk/pulls"
+        ]
+      },
+      {
+        "service": "ci",
+        "minRequests": 1,
+        "requiredPaths": [
+          "^/runs/"
+        ]
+      }
+    ]
+  },
+  "rubric": {
+    "artifact": {
+      "path": "ship-report.md",
+      "kind": "markdown"
+    },
+    "axes": [
+      {
+        "name": "safety",
+        "description": "Dry-run posture is explicit: no real credentials, clear stop-before-side-effect boundary."
+      },
+      {
+        "name": "mapping",
+        "description": "Fake endpoints/commands are mapped completely to the workflow steps."
+      },
+      {
+        "name": "signals",
+        "description": "Success, failure, retry, and abort conditions are all defined."
+      },
+      {
+        "name": "containment",
+        "description": "The plan cannot accidentally reach a real service from a mock context."
+      }
+    ]
+  },
+  "qualityFocus": [
+    "test CLI simulator",
+    "MCP simulator",
+    "branching gates",
+    "no real side effects"
+  ]
+}

package/data/craftbook-templates/so/social-digest/versions/1.0.1/test.json

@@ -77,8 +77,7 @@
       }
     ],
     "craftbookParams": {
-      "lookbackHours": "87600",
-      "workPath": "tasks/eval"
+      "lookbackHours": "87600"
     },
     "worker": {
       "name": "Nora",
@@ -91,19 +90,19 @@
     "checks": [
       {
         "kind": "minBytes",
-        "file": "tasks/eval/reports/social-digest-notes.md",
+        "file": "{{task.dir}}/reports/social-digest-notes.md",
         "bytes": 150,
         "artifact": true
       },
       {
         "kind": "minBytes",
-        "file": "tasks/eval/reports/social-digest-latest.md",
+        "file": "{{task.dir}}/reports/social-digest-latest.md",
         "bytes": 400,
         "artifact": true
       },
       {
         "kind": "contains",
-        "file": "tasks/eval/reports/social-digest-latest.md",
+        "file": "{{task.dir}}/reports/social-digest-latest.md",
         "pattern": "meike",
         "flags": "i",
         "label": "names the first mention author",
@@ -111,7 +110,7 @@
       },
       {
         "kind": "contains",
-        "file": "tasks/eval/reports/social-digest-latest.md",
+        "file": "{{task.dir}}/reports/social-digest-latest.md",
         "pattern": "stef-vandam|stef vandam",
         "flags": "i",
         "label": "names the second mention author",
@@ -119,14 +118,14 @@
       },
       {
         "kind": "contains",
-        "file": "tasks/eval/reports/social-digest-latest.md",
+        "file": "{{task.dir}}/reports/social-digest-latest.md",
         "pattern": "```gezel-action",
         "label": "carries at least one one-click action",
         "artifact": true
       },
       {
         "kind": "contains",
-        "file": "tasks/eval/reports/social-digest-latest.md",
+        "file": "{{task.dir}}/reports/social-digest-latest.md",
         "pattern": "meet-the-makers|Meet the makers",
         "label": "the standout post shows up in the digest",
         "artifact": true