@se-studio/contentful-cms 3.10.1 → 3.10.2

package/dist/generated/hosted-guide.d.ts

@@ -1,2 +1,2 @@
-export declare const HOSTED_GUIDE_TEMPLATE = "# cms-edit MCP guide\n\nUse this guide when editing Contentful draft content through a hosted MCP connector (Claude Desktop Integrations + OAuth, or Grok Build).\n\n{{SITE_INTRO}}\n\n## Safety rules\n\ncms-edit **cannot and will not**:\n\n- Publish or unpublish entries\n- Archive or delete published entries\n\nAll `save` operations create **drafts only**. A human must review and publish in Contentful.\n\n**Assets:** cms-edit **can upload** draft assets via **staged upload** (hosted) or `--url` (public HTTPS last resort). **`--base64` is not supported on hosted MCP.** Always **search and reuse existing assets first** (`index sync`, then `asset search` or `list --type media`). Use `--if-exists-by-filename` when uploading to avoid duplicates.\n\n## MCP tools\n\n| Tool | Purpose |\n|------|---------|\n| `cms_edit` | Run any cms-edit command via an `args` array |\n| `cms_edit_request_staged_upload` | Presigned URL for binaries. Raster/video: then `cms_edit_media` prepare. SVG/PDF/Lottie: then `asset upload --staged` |\n| `cms_edit_version` | Check the installed cms-edit version (no session required) |\n| `cms_edit_help` | Guide topics (`overview`, `workflow`, `commands`, `troubleshooting`, `hosted`) **or** playbooks (`create-from-json`, `fields-component`, \u2026) |\n\n## Customer editor pack (read first)\n\nWhen a site editor pack is deployed, these MCP resources provide **site-specific** context. Read them before creating or editing content:\n\n| Resource | Purpose |\n|----------|---------|\n| `cms-edit://customer/index` | List of all customer resources |\n| `cms-edit://customer/tasks-index` | Intent \u2192 task playbook map (filtered for this site) |\n| `cms-edit://customer/capabilities` | Machine-readable list of what this site supports |\n| `cms-edit://customer/checklist` | First-session onboarding steps |\n| `cms-edit://customer/overview` | Site name, space, staging URL, handoff rules |\n| `cms-edit://customer/routing` | Article/tag URL rules for this site |\n| `cms-edit://customer/production-site` | Production URL and how to fetch page markdown (do not guess `.md` URLs) |\n| `cms-edit://customer/components-index` | Component types and when to use each |\n| `cms-edit://customer/collections-index` | Collection types |\n| `cms-edit://customer/brand` | Brand voice and terminology (if configured) |\n| `cms-edit://customer/articles` | Article/publication playbook (if configured) |\n| `cms-edit://customer/defaults` | Featured image rules and article types (if configured) |\n| `cms-edit://customer/site-facts` | Optional canonical business facts (if configured) |\n| `cms-edit://customer/task-*` | Capability playbooks (e.g. `task-media-reuse-and-upload`) |\n\n**Workflow:** read `tasks-index` \u2192 `capabilities` \u2192 `checklist` \u2192 `overview` \u2192 `routing` \u2192 `production-site` \u2192 pick a `task-*` playbook for your goal \u2192 `components-index` \u2192 then use `cms_edit` commands.\n\n## MCP prompts\n\n| Prompt | Purpose |\n|--------|---------|\n| `cms_edit_verify` | Check version and connectivity |\n| `cms_edit_explore` | List site content and open a page |\n| `cms_edit_edit_page` | Edit a page through diff \u2192 save |\n| `cms_edit_edit_article` | Edit an article through diff \u2192 save |\n| `cms_edit_index_sync` | Build/refresh local index (assets, media, templates, taxonomy) |\n| `cms_edit_create_page` | Create a page with template selection |\n| `cms_edit_create_article` | Create an article (requires article type; featured image rules) |\n| `cms_edit_create_from_doc` | Create a page from a freeform document (infer \u2192 confirm \u2192 dry-run) |\n| `cms_edit_create_article_from_doc` | Create an article from a freeform document |\n| `cms_edit_taxonomy` | Import taxonomy entries (`create \u2026 --if-not-exists`) |\n| `cms_edit_bulk_seo` | Update SEO meta descriptions (read `task-seo-descriptions`) |\n| `cms_edit_bulk_find_replace` | Bulk find-replace in rich text across entries |\n| `cms_edit_bulk_article_tags` | Update tags on one or more articles |\n| `cms_edit_import_authors` | Create person/author entries from a list or sheet |\n| `cms_edit_media` | Find existing assets or upload new ones (reuse-first) |\n\nRead the full guide anytime via the `cms-edit://guide` resource or `cms_edit_help`.\n\n## End-to-end QA prompt\n\nFor release QA after cms-edit or editor-pack upgrades, use a **copy-paste Claude Desktop test script** (phases: connector \u2192 customer resources \u2192 discovery \u2192 session \u2192 schema \u2192 media \u2192 dry-run writes \u2192 negatives). See [e2e-test-prompt.md](./e2e-test-prompt.md) to author one per customer site. Example: Brightline `docs/cms-edit/MCP-E2E-TEST.md`.\n\n## Core workflow\n\nStandard editing flow: **open \u2192 snapshot \u2192 read \u2192 set/rtf \u2192 diff \u2192 save**\n\nEach step is a separate `cms_edit` call. Sessions persist in Redis between calls (hosted) or on disk (local). MCP auto-uses session `_mcp`.\n\n### 1. Open a page\n\n```\n[\"open\", \"--page-slug\", \"/\"]\n[\"open\", \"--page-slug\", \"/pricing\"]\n[\"open\", \"--id\", \"<entry-id>\"]\n[\"open\", \"--include-chrome\", \"--page-slug\", \"/pricing\"]\n```\n\nDefault `open` / `peek` fetch the page or article tree and the template entry, but **not** template menu/footer/navigation. Edit nav with `[\"nav\", \"open\", \"--nav-id\", \"<id>\"]` or `--nav-name` (not `--id`). Use `--include-chrome` only when you need those trees. Template `menu` is site-wide; `page.menu` overrides one page. Nested nav items are often shared \u2014 ask global vs page-specific before `set`.\n\n**Homepage (all SE Studio sites):** Contentful stores the marketing homepage slug as `index`. The live site URL is `/`. In cms-edit, `--page-slug /` and `--page-slug /index` are equivalent (both resolve to `index`). Read `cms-edit://customer/routing` for site-specific exceptions (e.g. Pedestal uses `home`).\n\n### 2. Snapshot \u2014 see the content tree\n\n```\n[\"snapshot\"]\n```\n\nOutput shows `@c0`, `@c1`, \u2026 refs for each component. Use these refs in `read`, `set`, and `rtf`.\n\n### 3. Read fields\n\n```\n[\"read\", \"@c1\"]\n[\"read\", \"@c1\", \"heading\"]\n```\n\n### 4. Edit fields\n\n```\n[\"set\", \"@c1\", \"heading\", \"New heading text\"]\n[\"rtf\", \"@c1\", \"body\", \"--content\", \"# Title\\n\\nParagraph text.\"]\n```\n\nFor rich text, prefer `--content` with inline markdown. Use `--base64` when special characters cause issues.\n\n### 5. Review and save\n\n```\n[\"diff\"]\n[\"save\"]\n```\n\nAlways run `diff` before `save`. To abandon changes: `[\"discard\"]`.\n\n## When to use Contentful UI\n\n| Task | Where |\n|------|-------|\n| Edit copy, add sections, create pages/articles | **Agent + cms-edit** (drafts) |\n| Upload new images and files | **Agent + cms-edit** \u2014 search/reuse first; raster/video via `cms_edit_media` import |\n| Publish approved drafts | **Contentful UI** |\n| Edit copy directly in Contentful | **Avoid** \u2014 use cms-edit instead |\n\ncms-edit saves **drafts only**. A colleague must **publish in Contentful** for changes to go live.\n\n## Asset upload (hosted MCP)\n\nHosted connectors have **no server filesystem** and **reject `asset upload --base64`**. Raster/video: staged intake then **media pipeline**. SVG/PDF/Lottie: staged then `asset upload --staged`.\n\n| Source | When | Example |\n|--------|------|---------|\n| **Staged + pipeline** | jpeg/png/webp/gif/mp4 | `cms_edit_media` prepare `stagedUploadIds` \u2192 import (required for video) |\n| **`--url`** | Pipeline-ready **images** only, or SVG/PDF | `[\"asset\", \"upload\", \"--url\", \"https://\u2026.blob.vercel-storage.com/ready.jpg\"]` |\n| **Staged consume** | SVG/PDF/Lottie | `asset upload --staged` |\n\n**After upload:** note the returned asset `id`, attach to the entry (`asset set @root media <id>` for person portraits), then `diff` \u2192 `save`. Upload alone does not update entries.\n\n**PNG** when transparency matters.\n\nFull workflow: `cms-edit://customer/task-media-reuse-and-upload`.\n\n### Common pitfalls (agents)\n\n| Issue | Do this |\n|-------|---------|\n| Singular Media on a component (`visual`) | `set @cN visual <mediaId> --link` \u2014 **not** `add \u2026 --target visual` |\n| Shared Drive folder has extra files | Import with `sourceNames` or `onlyWithOverrides` after catalog |\n| Prepare says no media / empty folder | Share folder **Viewer** with `media-pipeline-reader@se-media-pipeline.iam.gserviceaccount.com`, then re-prepare once |\n| MCP tool times out ~180s | `read` / `diff` / re-`catalog` before retrying the same write \u2014 server may already have saved |\n| `preview urls` has no Live Preview URL | Use staging + classic entry editor; row includes `unavailableReason` when CMA preview fails |\n| Reuse-first import matches old fileNames | dry-run first; set `overrides[].fileName` for unique names when duplication is required |\n\n### Staged upload\n\nKeeps binary **out of MCP tool args**. The agent uploads via `curl`, then passes only a short `uploadId`.\n\n1. `cms_edit_request_staged_upload` with `fileName`, `mimeType`, optional `byteLength`, optional `title`\n2. Run the returned `curlCommand` in a shell (POST multipart to `uploadUrl`)\n3. Follow `consumeArgs`: **media prepare** (`cms_edit_media` / `media prepare --staged-ids`) for jpeg/png/webp/gif/mp4; **asset upload --staged** for SVG/PDF/Lottie\n4. For pipeline jobs: wait \u2192 catalog \u2192 **import** (required for video). Then link and save\n\nHosted limit: **4 MB** per staged file (Vercel body limit). Rate limit: **1000 upload URLs per user per hour**. TTL: **60 minutes**. Raster/video staged files are intake (GET by the pipeline); they are not consumed into CMA.\n\n## Featured images (articles)\n\nRead `cms-edit://customer/defaults` when present. Otherwise:\n\n1. **Custom** \u2014 set `featuredImage` or `featuredImageAssetFilename` on the article when the editor provides a specific image\n2. **Article type** \u2014 if omitted, cms-edit inherits the linked **articleType** `featuredImage`\n3. **Site default** \u2014 from `editor.defaults` in the site config (also in `defaults.json` in the editor pack)\n\nRun `[\"index\", \"sync\"]` before resolving assets by filename.\n\n## JSON payloads (filesystem vs no filesystem)\n\nHosted MCP cannot see your disk. Same commands either way:\n\n- **Filesystem client:** write `page.json`, then `cms_edit [\"create\", \"from-json\", \"--json-base64\", \"<base64 of file>\", \"--dry-run\", \"--strict\"]`\n- **No filesystem (Desktop):** put the JSON in `--json` on the tool call\n\nNever use a local `cms-edit` binary, `--file`, or stdin for content.\n\n## Create content\n\n### Create a page\n\n```\n[\"schema\", \"component\"]\n[\"help\", \"create-from-json\"]\n[\"create\", \"from-json\", \"--json\", \"{...}\", \"--dry-run\", \"--strict\"]\n[\"create\", \"from-json\", \"--json\", \"{...}\"]\n```\n\nUse `\"components\"` (not `\"content\"`) and nested `\"type\"` (not `\"componentType\"`). Peek a live entry or `schema <ct>` before authoring field ids.\n\nEmpty `components` fails validation. Incremental `add`/`set` is for tiny patches only.\n\n### Create an article\n\nEvery article requires an **articleType**:\n\n```\n[\"index\", \"sync\"]\n[\"list\", \"--type\", \"articleType\"]\n[\"create\", \"article\", \"--slug\", \"resources/blog/my-post\", \"--title\", \"My Post\", \"--article-type-slug\", \"resources/blog\"]\n```\n\nOr `create from-json` with `\"type\": \"article\"` and `articleTypeSlug` / `articleTypeId`.\n\nUse MCP prompt `cms_edit_create_article` for guided creation.\n\n## Common sequences\n\n### List all pages\n\n```\n[\"list\", \"--type\", \"page\"]\n```\n\n### Refresh local index (assets, media, templates, taxonomy)\n\nRun before asset filename search, media-by-filename list, or fast template/taxonomy lookup:\n\n```\n[\"index\", \"sync\"]\n[\"index\", \"status\"]\n[\"list\", \"--type\", \"template\"]\n```\n\nAsset search and media-by-filename list **require** the index. Template/taxonomy list prefers the index when built.\n\n### Explore the homepage\n\n```\n[\"open\", \"--page-slug\", \"/\"]\n[\"snapshot\"]\n[\"read\", \"@c1\"]\n```\n\n### Edit copy on a page\n\n```\n[\"open\", \"--page-slug\", \"{{EXAMPLE_SLUG}}\"]\n[\"snapshot\"]\n[\"read\", \"@c1\", \"heading\"]\n[\"set\", \"@c1\", \"heading\", \"Updated heading\"]\n[\"diff\"]\n[\"save\"]\n```\n\n## Field discovery\n\n```\n[\"schema\", \"component\"]\n[\"list\", \"--type\", \"page\"]\n[\"search\", \"pricing\"]\n```\n\n## Grok Build setup\n\nGrok connects to the same hosted MCP URL as Claude. Add to `~/.grok/config.toml`:\n\n```toml\n[mcp_servers.cms-edit-brightline]\nurl = \"https://brightline.content.se.studio/api/mcp\"\nenabled = true\n```\n\nReplace the server name and host with your site's deployment. The URL **must** end with `/api/mcp` (not the site root).\n\n**Connect:**\n\n1. In Grok, run `/mcps` and select your cms-edit server\n2. Press **`i`** to initiate OAuth (not just enable the toggle)\n3. Sign in with Contentful when the browser opens\n4. Verify with `grok mcp doctor <server-name>` or call `cms_edit_version`\n\nOAuth is two steps: Grok obtains an MCP token, then you sign in with Contentful to link your account. If the browser flashes and closes immediately, reconnect after upgrading cms-edit (loopback redirect handling).\n\n## Hosted connector rules\n\n{{HOSTED_RULES}}\n\n## Session tips\n\n- After `open`, subsequent MCP commands share session `_mcp` automatically (stored in Redis on hosted).\n- Use `--session <name>` only when juggling multiple pages \u2014 pass it **before** the subcommand on every call: `[\"--session\", \"my-work\", \"open\", \"--page-slug\", \"/\"]`.\n- Snapshot JSON includes `sessionName` and `sessionKey` for debugging.\n- If you see \"No active session\", check `/api/health` (Redis must be ok), then run `open` again. Set `LOG_CMS_SESSION=1` on Vercel and search runtime logs for `[cms-edit:session]` (`redis save ok` / `redis load hit` / `requireSession miss`).\n\n### peek vs open vs read\n\n| Goal | MCP args |\n|------|----------|\n| Edit workflow (session) | `[\"open\", \"--page-slug\", \"/\"]` then `[\"read\", \"@c0\"]` |\n| Preview tree + fields, no session | `[\"peek\", \"--page-slug\", \"/pricing\"]` or `[\"peek\", \"--id\", \"<entry-id>\"]` (JSON includes `rootFields`) |\n| Read fields by entry ID, no session | `[\"read\", \"<entry-id>\", \"--id\"]` or `[\"read\", \"<entry-id>\", \"--id\", \"heading\"]` |\n| Debug session persistence | `[\"diagnose\"]` \u2014 reports store, userId, active session key |\n\n- **`peek` is one-shot** \u2014 prints tree and **root field values** (`rootFields` in JSON). No session is created; do not chain `read @ref` afterward.\n- **`open` / `nav open` persist** the session. `preview urls /slug` follows that slug even if leftover `_mcp` is another page (JSON may include `sessionRootIgnored`).\n- **Use flat lookup flags** for `open` and `peek` \u2014 same as CLI (`--page-slug`, `--id`, etc.). See `help lookup`. `nav open` uses `--nav-id` or `--nav-name` only.\n- **`read @root`** only works after `open` (uses the active session). The hint in peek output is for terminal users who already ran `open`.\n- **Do not invent flags** \u2014 there is no `--read`, `--then`, or chained peek+read syntax. One command per `cms_edit` call.\n\n## Troubleshooting\n\n| Problem | What to do |\n|---------|------------|\n| **401 / connection failed** | Reconnect: Claude \u2192 Settings \u2192 Integrations, or Grok `/mcps` \u2192 press `i`. Complete Contentful sign-in. If Grok says \u201CFailed to get tools after auth\u201D or `auth required`, check Vercel `[cms-edit:oauth] auth_failure`. `invalid_refresh_token` is a rotated/missing MCP refresh UUID (fixed in **3.6.1+** by not rotating). `missing_contentful_token` / `invalid_contentful_token` means the CMA cache is gone \u2014 **3.5.3+** returns `invalid_token` so Grok re-runs Contentful sign-in. If it still loops, delete that server\u2019s entry from `~/.grok/mcp_credentials.json` and press `i` again. Set `LOG_CMS_OAUTH=1` for connect success traces. |\n| **No cms_edit tool** | Confirm the integration is connected under Settings \u2192 Integrations |\n| **No active session** | Do not pass `--space` on hosted. Check `/api/health` \u2014 `redis` must be ok. MCP uses `_mcp` session automatically. Manual session: `[\"--session\", \"name\", \"open\", \"--page-slug\", \"/\"]`. Fallback: `[\"read\", \"<entry-id>\", \"--id\"]` |\n| **Wrong space** | Your Contentful account must have access to this deployment's space |\n| **Check version** | Call `cms_edit_version` or `curl /api/health` and read `cmsEdit` |\n\n## Articles\n\n**Recent articles** (by publication `date`, not last-updated):\n\n```json\n[\"list\", \"--type\", \"article\", \"--sort\", \"date\", \"-n\", \"10\"]\n```\n\nAlternative: `[\"search\", \"<keyword>\", \"--type\", \"article\"]` (also sorts by date).\n\n**Draft articles** \u2014 `peek` and `open` load drafts via CMA (not published-only). To discover drafts:\n\n```json\n[\"list\", \"--type\", \"article\", \"--sort\", \"date\"]\n[\"search\", \"<title>\", \"--type\", \"article\"]\n[\"list\", \"--type\", \"article\", \"--slug\", \"<slug>\"]\n```\n\nThen open: `[\"peek\", \"--article-slug\", \"<slug>\"]` or `[\"open\", \"--id\", \"<entry-id>\"]`. Use `--article-slug`, not `--page-slug`.\n\n**Published-only** is opt-in: `list --published` and `resolve --published` use the Delivery API index and omit never-published drafts.\n\nAfter upgrading cms-edit, run `[\"index\", \"sync\"]` once so the local index includes article `date` values (schema v4).\n\n**Index freshness on hosted MCP:** The catalog index is shared via Redis (gzipped SQLite blob). `index status` on a cold instance may show **not built** until `list`, `open`, or another index command runs \u2014 that is normal. Committed writes (`save`, `create page`, `asset upload`, \u2026) mark the index dirty; the next discovery command refreshes it. In-session `set` edits do not trigger a sync.\n\n## Command reference (cms_edit args)\n\n**Discovery:** `list --type <type>`, `search <query>`, `schema <contentType>`, `index sync`, `index status`\n\n**Reading:** `open --page-slug <slug>`, `open --id <id>`, `peek --page-slug <slug>`, `peek --id <id>`, `snapshot`, `read <ref>`, `read <ref> <field>`, `read <id> --id [field]`\n\n**Diagnostics:** `diagnose` \u2014 session store, userId, active session (after open/read issues)\n\n**Editing:** `set <ref> <field> <value>`, `rtf <ref> <field> --content \"<md>\"`, `diff`, `save`, `discard`\n\n**Creating:** `create from-json --json \u2026 --dry-run --strict`, `export from-json`, `apply from-json --json \u2026`, `create page` / `create article` for empty shells, `add` for tiny inserts\n\n**Other:** `health`, `version`, `guide [topic]`\n";
+export declare const HOSTED_GUIDE_TEMPLATE = "# cms-edit MCP guide\n\nUse this guide when editing Contentful draft content through a hosted MCP connector (Claude Desktop Integrations + OAuth, or Grok Build).\n\n{{SITE_INTRO}}\n\n## Safety rules\n\ncms-edit **cannot and will not**:\n\n- Publish or unpublish entries\n- Archive or delete published entries\n\nAll `save` operations create **drafts only**. A human must review and publish in Contentful.\n\n**Assets:** cms-edit **can upload** draft assets via **staged upload** (hosted) or `--url` (public HTTPS last resort). **`--base64` is not supported on hosted MCP.** Always **search and reuse existing assets first** (`index sync`, then `asset search` or `list --type media`). Use `--if-exists-by-filename` when uploading to avoid duplicates.\n\n## MCP tools\n\n| Tool | Purpose |\n|------|---------|\n| `cms_edit` | Run any cms-edit command via an `args` array |\n| `cms_edit_request_staged_upload` | Presigned URL for binaries. Raster/video: then `cms_edit_media` prepare. SVG/PDF/Lottie: then `asset upload --staged` |\n| `cms_edit_version` | Check the installed cms-edit version (no session required) |\n| `cms_edit_help` | Guide topics (`overview`, `workflow`, `commands`, `troubleshooting`, `hosted`) **or** playbooks (`create-from-json`, `fields-html-component`, `fields-component`, \u2026) |\n\n## Customer editor pack (read first)\n\nWhen a site editor pack is deployed, these MCP resources provide **site-specific** context. Read them before creating or editing content:\n\n| Resource | Purpose |\n|----------|---------|\n| `cms-edit://customer/index` | List of all customer resources |\n| `cms-edit://customer/tasks-index` | Intent \u2192 task playbook map (filtered for this site) |\n| `cms-edit://customer/capabilities` | Machine-readable list of what this site supports |\n| `cms-edit://customer/checklist` | First-session onboarding steps |\n| `cms-edit://customer/overview` | Site name, space, staging URL, handoff rules |\n| `cms-edit://customer/routing` | Article/tag URL rules for this site |\n| `cms-edit://customer/production-site` | Production URL and how to fetch page markdown (do not guess `.md` URLs) |\n| `cms-edit://customer/components-index` | Component types and when to use each |\n| `cms-edit://customer/collections-index` | Collection types |\n| `cms-edit://customer/brand` | Brand voice and terminology (if configured) |\n| `cms-edit://customer/articles` | Article/publication playbook (if configured) |\n| `cms-edit://customer/defaults` | Featured image rules and article types (if configured) |\n| `cms-edit://customer/site-facts` | Optional canonical business facts (if configured) |\n| `cms-edit://customer/task-*` | Capability playbooks (e.g. `task-media-reuse-and-upload`) |\n\n**Workflow:** read `tasks-index` \u2192 `capabilities` \u2192 `checklist` \u2192 `overview` \u2192 `routing` \u2192 `production-site` \u2192 pick a `task-*` playbook for your goal \u2192 `components-index` \u2192 then use `cms_edit` commands.\n\n## MCP prompts\n\n| Prompt | Purpose |\n|--------|---------|\n| `cms_edit_verify` | Check version and connectivity |\n| `cms_edit_explore` | List site content and open a page |\n| `cms_edit_edit_page` | Edit a page through diff \u2192 save |\n| `cms_edit_edit_article` | Edit an article through diff \u2192 save |\n| `cms_edit_index_sync` | Build/refresh local index (assets, media, templates, taxonomy) |\n| `cms_edit_create_page` | Create a page with template selection |\n| `cms_edit_create_article` | Create an article (requires article type; featured image rules) |\n| `cms_edit_create_from_doc` | Create a page from a freeform document (infer \u2192 confirm \u2192 dry-run) |\n| `cms_edit_create_article_from_doc` | Create an article from a freeform document |\n| `cms_edit_taxonomy` | Import taxonomy entries (`create \u2026 --if-not-exists`) |\n| `cms_edit_bulk_seo` | Update SEO meta descriptions (read `task-seo-descriptions`) |\n| `cms_edit_bulk_find_replace` | Bulk find-replace in rich text across entries |\n| `cms_edit_bulk_article_tags` | Update tags on one or more articles |\n| `cms_edit_import_authors` | Create person/author entries from a list or sheet |\n| `cms_edit_media` | Find existing assets or upload new ones (reuse-first) |\n\nRead the full guide anytime via the `cms-edit://guide` resource or `cms_edit_help`.\n\n## End-to-end QA prompt\n\nFor release QA after cms-edit or editor-pack upgrades, use a **copy-paste Claude Desktop test script** (phases: connector \u2192 customer resources \u2192 discovery \u2192 session \u2192 schema \u2192 media \u2192 dry-run writes \u2192 negatives). See [e2e-test-prompt.md](./e2e-test-prompt.md) to author one per customer site. Example: Brightline `docs/cms-edit/MCP-E2E-TEST.md`.\n\n## Core workflow\n\nStandard editing flow: **open \u2192 snapshot \u2192 read \u2192 set/rtf \u2192 diff \u2192 save**\n\nEach step is a separate `cms_edit` call. Sessions persist in Redis between calls (hosted) or on disk (local). MCP auto-uses session `_mcp`.\n\n### 1. Open a page\n\n```\n[\"open\", \"--page-slug\", \"/\"]\n[\"open\", \"--page-slug\", \"/pricing\"]\n[\"open\", \"--id\", \"<entry-id>\"]\n[\"open\", \"--include-chrome\", \"--page-slug\", \"/pricing\"]\n```\n\nDefault `open` / `peek` fetch the page or article tree and the template entry, but **not** template menu/footer/navigation. Edit nav with `[\"nav\", \"open\", \"--nav-id\", \"<id>\"]` or `--nav-name` (not `--id`). Use `--include-chrome` only when you need those trees. Template `menu` is site-wide; `page.menu` overrides one page. Nested nav items are often shared \u2014 ask global vs page-specific before `set`.\n\n**Homepage (all SE Studio sites):** Contentful stores the marketing homepage slug as `index`. The live site URL is `/`. In cms-edit, `--page-slug /` and `--page-slug /index` are equivalent (both resolve to `index`). Read `cms-edit://customer/routing` for site-specific exceptions (e.g. Pedestal uses `home`).\n\n### 2. Snapshot \u2014 see the content tree\n\n```\n[\"snapshot\"]\n```\n\nOutput shows `@c0`, `@c1`, \u2026 refs for each component. Use these refs in `read`, `set`, and `rtf`.\n\n### 3. Read fields\n\n```\n[\"read\", \"@c1\"]\n[\"read\", \"@c1\", \"heading\"]\n```\n\n### 4. Edit fields\n\n```\n[\"set\", \"@c1\", \"heading\", \"New heading text\"]\n[\"rtf\", \"@c1\", \"body\", \"--content\", \"# Title\\n\\nParagraph text.\"]\n```\n\nFor rich text, prefer `--content` with inline markdown. Use `--base64` when special characters cause issues.\n\n### 5. Review and save\n\n```\n[\"diff\"]\n[\"save\"]\n```\n\nAlways run `diff` before `save`. To abandon changes: `[\"discard\"]`.\n\n## When to use Contentful UI\n\n| Task | Where |\n|------|-------|\n| Edit copy, add sections, create pages/articles | **Agent + cms-edit** (drafts) |\n| Upload new images and files | **Agent + cms-edit** \u2014 search/reuse first; raster/video via `cms_edit_media` import |\n| Publish approved drafts | **Contentful UI** |\n| Edit copy directly in Contentful | **Avoid** \u2014 use cms-edit instead |\n\ncms-edit saves **drafts only**. A colleague must **publish in Contentful** for changes to go live.\n\n## Asset upload (hosted MCP)\n\nHosted connectors have **no server filesystem** and **reject `asset upload --base64`**. Raster/video: staged intake then **media pipeline**. SVG/PDF/Lottie: staged then `asset upload --staged`.\n\n| Source | When | Example |\n|--------|------|---------|\n| **Staged + pipeline** | jpeg/png/webp/gif/mp4 | `cms_edit_media` prepare `stagedUploadIds` \u2192 import (required for video) |\n| **`--url`** | Pipeline-ready **images** only, or SVG/PDF | `[\"asset\", \"upload\", \"--url\", \"https://\u2026.blob.vercel-storage.com/ready.jpg\"]` |\n| **Staged consume** | SVG/PDF/Lottie | `asset upload --staged` |\n\n**After upload:** note the returned asset `id`, attach to the entry (`asset set @root media <id>` for person portraits), then `diff` \u2192 `save`. Upload alone does not update entries.\n\n**PNG** when transparency matters.\n\nFull workflow: `cms-edit://customer/task-media-reuse-and-upload`.\n\n### Common pitfalls (agents)\n\n| Issue | Do this |\n|-------|---------|\n| Singular Media on a component (`visual`) | `set @cN visual <mediaId> --link` \u2014 **not** `add \u2026 --target visual` |\n| Shared Drive folder has extra files | Import with `sourceNames` or `onlyWithOverrides` after catalog |\n| Prepare says no media / empty folder | Share folder **Viewer** with `media-pipeline-reader@se-media-pipeline.iam.gserviceaccount.com`, then re-prepare once |\n| MCP tool times out ~180s | `read` / `diff` / re-`catalog` before retrying the same write \u2014 server may already have saved |\n| `preview urls` has no Live Preview URL | Use staging + classic entry editor; row includes `unavailableReason` when CMA preview fails |\n| Reuse-first import matches old fileNames | dry-run first; set `overrides[].fileName` for unique names when duplication is required |\n\n### Staged upload\n\nKeeps binary **out of MCP tool args**. The agent uploads via `curl`, then passes only a short `uploadId`.\n\n1. `cms_edit_request_staged_upload` with `fileName`, `mimeType`, optional `byteLength`, optional `title`\n2. Run the returned `curlCommand` in a shell (POST multipart to `uploadUrl`)\n3. Follow `consumeArgs`: **media prepare** (`cms_edit_media` / `media prepare --staged-ids`) for jpeg/png/webp/gif/mp4; **asset upload --staged** for SVG/PDF/Lottie\n4. For pipeline jobs: wait \u2192 catalog \u2192 **import** (required for video). Then link and save\n\nHosted limit: **4 MB** per staged file (Vercel body limit). Rate limit: **1000 upload URLs per user per hour**. TTL: **60 minutes**. Raster/video staged files are intake (GET by the pipeline); they are not consumed into CMA.\n\n## Featured images (articles)\n\nRead `cms-edit://customer/defaults` when present. Otherwise:\n\n1. **Custom** \u2014 set `featuredImage` or `featuredImageAssetFilename` on the article when the editor provides a specific image\n2. **Article type** \u2014 if omitted, cms-edit inherits the linked **articleType** `featuredImage`\n3. **Site default** \u2014 from `editor.defaults` in the site config (also in `defaults.json` in the editor pack)\n\nRun `[\"index\", \"sync\"]` before resolving assets by filename.\n\n## JSON payloads (filesystem vs no filesystem)\n\nHosted MCP cannot see your disk. Same commands either way:\n\n- **Filesystem client:** write `page.json`, then `cms_edit [\"create\", \"from-json\", \"--json-base64\", \"<base64 of file>\", \"--dry-run\", \"--strict\"]`\n- **No filesystem (Desktop):** put the JSON in `--json` on the tool call\n\nNever use a local `cms-edit` binary, `--file`, or stdin for content.\n\n## Create content\n\n### Create a page\n\n```\n[\"schema\", \"component\"]\n[\"help\", \"create-from-json\"]\n[\"create\", \"from-json\", \"--json\", \"{...}\", \"--dry-run\", \"--strict\"]\n[\"create\", \"from-json\", \"--json\", \"{...}\"]\n```\n\nUse `\"components\"` (not `\"content\"`) and nested `\"type\"` (not `\"componentType\"`). Peek a live entry or `schema <ct>` before authoring field ids.\n\nEmpty `components` fails validation. Incremental `add`/`set` is for tiny patches only.\n\n### Create an article\n\nEvery article requires an **articleType**:\n\n```\n[\"index\", \"sync\"]\n[\"list\", \"--type\", \"articleType\"]\n[\"create\", \"article\", \"--slug\", \"resources/blog/my-post\", \"--title\", \"My Post\", \"--article-type-slug\", \"resources/blog\"]\n```\n\nOr `create from-json` with `\"type\": \"article\"` and `articleTypeSlug` / `articleTypeId`.\n\nUse MCP prompt `cms_edit_create_article` for guided creation.\n\n## Common sequences\n\n### List all pages\n\n```\n[\"list\", \"--type\", \"page\"]\n```\n\n### Refresh local index (assets, media, templates, taxonomy)\n\nRun before asset filename search, media-by-filename list, or fast template/taxonomy lookup:\n\n```\n[\"index\", \"sync\"]\n[\"index\", \"status\"]\n[\"list\", \"--type\", \"template\"]\n```\n\nAsset search and media-by-filename list **require** the index. Template/taxonomy list prefers the index when built.\n\n### Explore the homepage\n\n```\n[\"open\", \"--page-slug\", \"/\"]\n[\"snapshot\"]\n[\"read\", \"@c1\"]\n```\n\n### Edit copy on a page\n\n```\n[\"open\", \"--page-slug\", \"{{EXAMPLE_SLUG}}\"]\n[\"snapshot\"]\n[\"read\", \"@c1\", \"heading\"]\n[\"set\", \"@c1\", \"heading\", \"Updated heading\"]\n[\"diff\"]\n[\"save\"]\n```\n\n## Field discovery\n\n```\n[\"schema\", \"component\"]\n[\"list\", \"--type\", \"page\"]\n[\"search\", \"pricing\"]\n```\n\n## Grok Build setup\n\nGrok connects to the same hosted MCP URL as Claude. Add to `~/.grok/config.toml`:\n\n```toml\n[mcp_servers.cms-edit-brightline]\nurl = \"https://brightline.content.se.studio/api/mcp\"\nenabled = true\n```\n\nReplace the server name and host with your site's deployment. The URL **must** end with `/api/mcp` (not the site root).\n\n**Connect:**\n\n1. In Grok, run `/mcps` and select your cms-edit server\n2. Press **`i`** to initiate OAuth (not just enable the toggle)\n3. Sign in with Contentful when the browser opens\n4. Verify with `grok mcp doctor <server-name>` or call `cms_edit_version`\n\nOAuth is two steps: Grok obtains an MCP token, then you sign in with Contentful to link your account. If the browser flashes and closes immediately, reconnect after upgrading cms-edit (loopback redirect handling).\n\n## Hosted connector rules\n\n{{HOSTED_RULES}}\n\n## Session tips\n\n- After `open`, subsequent MCP commands share session `_mcp` automatically (stored in Redis on hosted).\n- Use `--session <name>` only when juggling multiple pages \u2014 pass it **before** the subcommand on every call: `[\"--session\", \"my-work\", \"open\", \"--page-slug\", \"/\"]`.\n- Snapshot JSON includes `sessionName` and `sessionKey` for debugging.\n- If you see \"No active session\", check `/api/health` (Redis must be ok), then run `open` again. Set `LOG_CMS_SESSION=1` on Vercel and search runtime logs for `[cms-edit:session]` (`redis save ok` / `redis load hit` / `requireSession miss`).\n\n### peek vs open vs read\n\n| Goal | MCP args |\n|------|----------|\n| Edit workflow (session) | `[\"open\", \"--page-slug\", \"/\"]` then `[\"read\", \"@c0\"]` |\n| Preview tree + fields, no session | `[\"peek\", \"--page-slug\", \"/pricing\"]` or `[\"peek\", \"--id\", \"<entry-id>\"]` (JSON includes `rootFields`) |\n| Read fields by entry ID, no session | `[\"read\", \"<entry-id>\", \"--id\"]` or `[\"read\", \"<entry-id>\", \"--id\", \"heading\"]` |\n| Debug session persistence | `[\"diagnose\"]` \u2014 reports store, userId, active session key |\n\n- **`peek` is one-shot** \u2014 prints tree and **root field values** (`rootFields` in JSON). No session is created; do not chain `read @ref` afterward.\n- **`open` / `nav open` persist** the session. `preview urls /slug` follows that slug even if leftover `_mcp` is another page (JSON may include `sessionRootIgnored`).\n- **Use flat lookup flags** for `open` and `peek` \u2014 same as CLI (`--page-slug`, `--id`, etc.). See `help lookup`. `nav open` uses `--nav-id` or `--nav-name` only.\n- **`read @root`** only works after `open` (uses the active session). The hint in peek output is for terminal users who already ran `open`.\n- **Do not invent flags** \u2014 there is no `--read`, `--then`, or chained peek+read syntax. One command per `cms_edit` call.\n\n## Troubleshooting\n\n| Problem | What to do |\n|---------|------------|\n| **401 / connection failed** | Reconnect: Claude \u2192 Settings \u2192 Integrations, or Grok `/mcps` \u2192 press `i`. Complete Contentful sign-in. If Grok says \u201CFailed to get tools after auth\u201D or `auth required`, check Vercel `[cms-edit:oauth] auth_failure`. `invalid_refresh_token` is a rotated/missing MCP refresh UUID (fixed in **3.6.1+** by not rotating). `missing_contentful_token` / `invalid_contentful_token` means the CMA cache is gone \u2014 **3.5.3+** returns `invalid_token` so Grok re-runs Contentful sign-in. If it still loops, delete that server\u2019s entry from `~/.grok/mcp_credentials.json` and press `i` again. Set `LOG_CMS_OAUTH=1` for connect success traces. |\n| **No cms_edit tool** | Confirm the integration is connected under Settings \u2192 Integrations |\n| **No active session** | Do not pass `--space` on hosted. Check `/api/health` \u2014 `redis` must be ok. MCP uses `_mcp` session automatically. Manual session: `[\"--session\", \"name\", \"open\", \"--page-slug\", \"/\"]`. Fallback: `[\"read\", \"<entry-id>\", \"--id\"]` |\n| **Wrong space** | Your Contentful account must have access to this deployment's space |\n| **Check version** | Call `cms_edit_version` or `curl /api/health` and read `cmsEdit` |\n\n## Articles\n\n**Recent articles** (by publication `date`, not last-updated):\n\n```json\n[\"list\", \"--type\", \"article\", \"--sort\", \"date\", \"-n\", \"10\"]\n```\n\nAlternative: `[\"search\", \"<keyword>\", \"--type\", \"article\"]` (also sorts by date).\n\n**Draft articles** \u2014 `peek` and `open` load drafts via CMA (not published-only). To discover drafts:\n\n```json\n[\"list\", \"--type\", \"article\", \"--sort\", \"date\"]\n[\"search\", \"<title>\", \"--type\", \"article\"]\n[\"list\", \"--type\", \"article\", \"--slug\", \"<slug>\"]\n```\n\nThen open: `[\"peek\", \"--article-slug\", \"<slug>\"]` or `[\"open\", \"--id\", \"<entry-id>\"]`. Use `--article-slug`, not `--page-slug`.\n\n**Published-only** is opt-in: `list --published` and `resolve --published` use the Delivery API index and omit never-published drafts.\n\nAfter upgrading cms-edit, run `[\"index\", \"sync\"]` once so the local index includes article `date` values (schema v4).\n\n**Index freshness on hosted MCP:** The catalog index is shared via Redis (gzipped SQLite blob). `index status` on a cold instance may show **not built** until `list`, `open`, or another index command runs \u2014 that is normal. Committed writes (`save`, `create page`, `asset upload`, \u2026) mark the index dirty; the next discovery command refreshes it. In-session `set` edits do not trigger a sync.\n\n## Command reference (cms_edit args)\n\n**Discovery:** `list --type <type>`, `search <query>`, `schema <contentType>`, `index sync`, `index status`\n\n**Reading:** `open --page-slug <slug>`, `open --id <id>`, `peek --page-slug <slug>`, `peek --id <id>`, `snapshot`, `read <ref>`, `read <ref> <field>`, `read <id> --id [field]`\n\n**Diagnostics:** `diagnose` \u2014 session store, userId, active session (after open/read issues)\n\n**Editing:** `set <ref> <field> <value>`, `rtf <ref> <field> --content \"<md>\"`, `diff`, `save`, `discard`\n\n**Creating:** `create from-json --json \u2026 --dry-run --strict`, `export from-json`, `apply from-json --json \u2026`, `create page` / `create article` for empty shells, `add` for tiny inserts\n\n**Other:** `health`, `version`, `guide [topic]`\n";
 //# sourceMappingURL=hosted-guide.d.ts.map

package/dist/generated/hosted-guide.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"hosted-guide.d.ts","sourceRoot":"","sources":["../../src/generated/hosted-guide.ts"],"names":[],"mappings":"AACA,eAAO,MAAM,qBAAqB,k6kBAAumkB,CAAC"}
+{"version":3,"file":"hosted-guide.d.ts","sourceRoot":"","sources":["../../src/generated/hosted-guide.ts"],"names":[],"mappings":"AACA,eAAO,MAAM,qBAAqB,27kBAAgokB,CAAC"}

package/dist/generated/hosted-guide.js

@@ -1,3 +1,3 @@
 // Generated by scripts/generate-embedded-assets.mjs — do not edit manually.
-export const HOSTED_GUIDE_TEMPLATE = "# cms-edit MCP guide\n\nUse this guide when editing Contentful draft content through a hosted MCP connector (Claude Desktop Integrations + OAuth, or Grok Build).\n\n{{SITE_INTRO}}\n\n## Safety rules\n\ncms-edit **cannot and will not**:\n\n- Publish or unpublish entries\n- Archive or delete published entries\n\nAll `save` operations create **drafts only**. A human must review and publish in Contentful.\n\n**Assets:** cms-edit **can upload** draft assets via **staged upload** (hosted) or `--url` (public HTTPS last resort). **`--base64` is not supported on hosted MCP.** Always **search and reuse existing assets first** (`index sync`, then `asset search` or `list --type media`). Use `--if-exists-by-filename` when uploading to avoid duplicates.\n\n## MCP tools\n\n| Tool | Purpose |\n|------|---------|\n| `cms_edit` | Run any cms-edit command via an `args` array |\n| `cms_edit_request_staged_upload` | Presigned URL for binaries. Raster/video: then `cms_edit_media` prepare. SVG/PDF/Lottie: then `asset upload --staged` |\n| `cms_edit_version` | Check the installed cms-edit version (no session required) |\n| `cms_edit_help` | Guide topics (`overview`, `workflow`, `commands`, `troubleshooting`, `hosted`) **or** playbooks (`create-from-json`, `fields-component`, …) |\n\n## Customer editor pack (read first)\n\nWhen a site editor pack is deployed, these MCP resources provide **site-specific** context. Read them before creating or editing content:\n\n| Resource | Purpose |\n|----------|---------|\n| `cms-edit://customer/index` | List of all customer resources |\n| `cms-edit://customer/tasks-index` | Intent → task playbook map (filtered for this site) |\n| `cms-edit://customer/capabilities` | Machine-readable list of what this site supports |\n| `cms-edit://customer/checklist` | First-session onboarding steps |\n| `cms-edit://customer/overview` | Site name, space, staging URL, handoff rules |\n| `cms-edit://customer/routing` | Article/tag URL rules for this site |\n| `cms-edit://customer/production-site` | Production URL and how to fetch page markdown (do not guess `.md` URLs) |\n| `cms-edit://customer/components-index` | Component types and when to use each |\n| `cms-edit://customer/collections-index` | Collection types |\n| `cms-edit://customer/brand` | Brand voice and terminology (if configured) |\n| `cms-edit://customer/articles` | Article/publication playbook (if configured) |\n| `cms-edit://customer/defaults` | Featured image rules and article types (if configured) |\n| `cms-edit://customer/site-facts` | Optional canonical business facts (if configured) |\n| `cms-edit://customer/task-*` | Capability playbooks (e.g. `task-media-reuse-and-upload`) |\n\n**Workflow:** read `tasks-index` → `capabilities` → `checklist` → `overview` → `routing` → `production-site` → pick a `task-*` playbook for your goal → `components-index` → then use `cms_edit` commands.\n\n## MCP prompts\n\n| Prompt | Purpose |\n|--------|---------|\n| `cms_edit_verify` | Check version and connectivity |\n| `cms_edit_explore` | List site content and open a page |\n| `cms_edit_edit_page` | Edit a page through diff → save |\n| `cms_edit_edit_article` | Edit an article through diff → save |\n| `cms_edit_index_sync` | Build/refresh local index (assets, media, templates, taxonomy) |\n| `cms_edit_create_page` | Create a page with template selection |\n| `cms_edit_create_article` | Create an article (requires article type; featured image rules) |\n| `cms_edit_create_from_doc` | Create a page from a freeform document (infer → confirm → dry-run) |\n| `cms_edit_create_article_from_doc` | Create an article from a freeform document |\n| `cms_edit_taxonomy` | Import taxonomy entries (`create … --if-not-exists`) |\n| `cms_edit_bulk_seo` | Update SEO meta descriptions (read `task-seo-descriptions`) |\n| `cms_edit_bulk_find_replace` | Bulk find-replace in rich text across entries |\n| `cms_edit_bulk_article_tags` | Update tags on one or more articles |\n| `cms_edit_import_authors` | Create person/author entries from a list or sheet |\n| `cms_edit_media` | Find existing assets or upload new ones (reuse-first) |\n\nRead the full guide anytime via the `cms-edit://guide` resource or `cms_edit_help`.\n\n## End-to-end QA prompt\n\nFor release QA after cms-edit or editor-pack upgrades, use a **copy-paste Claude Desktop test script** (phases: connector → customer resources → discovery → session → schema → media → dry-run writes → negatives). See [e2e-test-prompt.md](./e2e-test-prompt.md) to author one per customer site. Example: Brightline `docs/cms-edit/MCP-E2E-TEST.md`.\n\n## Core workflow\n\nStandard editing flow: **open → snapshot → read → set/rtf → diff → save**\n\nEach step is a separate `cms_edit` call. Sessions persist in Redis between calls (hosted) or on disk (local). MCP auto-uses session `_mcp`.\n\n### 1. Open a page\n\n```\n[\"open\", \"--page-slug\", \"/\"]\n[\"open\", \"--page-slug\", \"/pricing\"]\n[\"open\", \"--id\", \"<entry-id>\"]\n[\"open\", \"--include-chrome\", \"--page-slug\", \"/pricing\"]\n```\n\nDefault `open` / `peek` fetch the page or article tree and the template entry, but **not** template menu/footer/navigation. Edit nav with `[\"nav\", \"open\", \"--nav-id\", \"<id>\"]` or `--nav-name` (not `--id`). Use `--include-chrome` only when you need those trees. Template `menu` is site-wide; `page.menu` overrides one page. Nested nav items are often shared — ask global vs page-specific before `set`.\n\n**Homepage (all SE Studio sites):** Contentful stores the marketing homepage slug as `index`. The live site URL is `/`. In cms-edit, `--page-slug /` and `--page-slug /index` are equivalent (both resolve to `index`). Read `cms-edit://customer/routing` for site-specific exceptions (e.g. Pedestal uses `home`).\n\n### 2. Snapshot — see the content tree\n\n```\n[\"snapshot\"]\n```\n\nOutput shows `@c0`, `@c1`, … refs for each component. Use these refs in `read`, `set`, and `rtf`.\n\n### 3. Read fields\n\n```\n[\"read\", \"@c1\"]\n[\"read\", \"@c1\", \"heading\"]\n```\n\n### 4. Edit fields\n\n```\n[\"set\", \"@c1\", \"heading\", \"New heading text\"]\n[\"rtf\", \"@c1\", \"body\", \"--content\", \"# Title\\n\\nParagraph text.\"]\n```\n\nFor rich text, prefer `--content` with inline markdown. Use `--base64` when special characters cause issues.\n\n### 5. Review and save\n\n```\n[\"diff\"]\n[\"save\"]\n```\n\nAlways run `diff` before `save`. To abandon changes: `[\"discard\"]`.\n\n## When to use Contentful UI\n\n| Task | Where |\n|------|-------|\n| Edit copy, add sections, create pages/articles | **Agent + cms-edit** (drafts) |\n| Upload new images and files | **Agent + cms-edit** — search/reuse first; raster/video via `cms_edit_media` import |\n| Publish approved drafts | **Contentful UI** |\n| Edit copy directly in Contentful | **Avoid** — use cms-edit instead |\n\ncms-edit saves **drafts only**. A colleague must **publish in Contentful** for changes to go live.\n\n## Asset upload (hosted MCP)\n\nHosted connectors have **no server filesystem** and **reject `asset upload --base64`**. Raster/video: staged intake then **media pipeline**. SVG/PDF/Lottie: staged then `asset upload --staged`.\n\n| Source | When | Example |\n|--------|------|---------|\n| **Staged + pipeline** | jpeg/png/webp/gif/mp4 | `cms_edit_media` prepare `stagedUploadIds` → import (required for video) |\n| **`--url`** | Pipeline-ready **images** only, or SVG/PDF | `[\"asset\", \"upload\", \"--url\", \"https://….blob.vercel-storage.com/ready.jpg\"]` |\n| **Staged consume** | SVG/PDF/Lottie | `asset upload --staged` |\n\n**After upload:** note the returned asset `id`, attach to the entry (`asset set @root media <id>` for person portraits), then `diff` → `save`. Upload alone does not update entries.\n\n**PNG** when transparency matters.\n\nFull workflow: `cms-edit://customer/task-media-reuse-and-upload`.\n\n### Common pitfalls (agents)\n\n| Issue | Do this |\n|-------|---------|\n| Singular Media on a component (`visual`) | `set @cN visual <mediaId> --link` — **not** `add … --target visual` |\n| Shared Drive folder has extra files | Import with `sourceNames` or `onlyWithOverrides` after catalog |\n| Prepare says no media / empty folder | Share folder **Viewer** with `media-pipeline-reader@se-media-pipeline.iam.gserviceaccount.com`, then re-prepare once |\n| MCP tool times out ~180s | `read` / `diff` / re-`catalog` before retrying the same write — server may already have saved |\n| `preview urls` has no Live Preview URL | Use staging + classic entry editor; row includes `unavailableReason` when CMA preview fails |\n| Reuse-first import matches old fileNames | dry-run first; set `overrides[].fileName` for unique names when duplication is required |\n\n### Staged upload\n\nKeeps binary **out of MCP tool args**. The agent uploads via `curl`, then passes only a short `uploadId`.\n\n1. `cms_edit_request_staged_upload` with `fileName`, `mimeType`, optional `byteLength`, optional `title`\n2. Run the returned `curlCommand` in a shell (POST multipart to `uploadUrl`)\n3. Follow `consumeArgs`: **media prepare** (`cms_edit_media` / `media prepare --staged-ids`) for jpeg/png/webp/gif/mp4; **asset upload --staged** for SVG/PDF/Lottie\n4. For pipeline jobs: wait → catalog → **import** (required for video). Then link and save\n\nHosted limit: **4 MB** per staged file (Vercel body limit). Rate limit: **1000 upload URLs per user per hour**. TTL: **60 minutes**. Raster/video staged files are intake (GET by the pipeline); they are not consumed into CMA.\n\n## Featured images (articles)\n\nRead `cms-edit://customer/defaults` when present. Otherwise:\n\n1. **Custom** — set `featuredImage` or `featuredImageAssetFilename` on the article when the editor provides a specific image\n2. **Article type** — if omitted, cms-edit inherits the linked **articleType** `featuredImage`\n3. **Site default** — from `editor.defaults` in the site config (also in `defaults.json` in the editor pack)\n\nRun `[\"index\", \"sync\"]` before resolving assets by filename.\n\n## JSON payloads (filesystem vs no filesystem)\n\nHosted MCP cannot see your disk. Same commands either way:\n\n- **Filesystem client:** write `page.json`, then `cms_edit [\"create\", \"from-json\", \"--json-base64\", \"<base64 of file>\", \"--dry-run\", \"--strict\"]`\n- **No filesystem (Desktop):** put the JSON in `--json` on the tool call\n\nNever use a local `cms-edit` binary, `--file`, or stdin for content.\n\n## Create content\n\n### Create a page\n\n```\n[\"schema\", \"component\"]\n[\"help\", \"create-from-json\"]\n[\"create\", \"from-json\", \"--json\", \"{...}\", \"--dry-run\", \"--strict\"]\n[\"create\", \"from-json\", \"--json\", \"{...}\"]\n```\n\nUse `\"components\"` (not `\"content\"`) and nested `\"type\"` (not `\"componentType\"`). Peek a live entry or `schema <ct>` before authoring field ids.\n\nEmpty `components` fails validation. Incremental `add`/`set` is for tiny patches only.\n\n### Create an article\n\nEvery article requires an **articleType**:\n\n```\n[\"index\", \"sync\"]\n[\"list\", \"--type\", \"articleType\"]\n[\"create\", \"article\", \"--slug\", \"resources/blog/my-post\", \"--title\", \"My Post\", \"--article-type-slug\", \"resources/blog\"]\n```\n\nOr `create from-json` with `\"type\": \"article\"` and `articleTypeSlug` / `articleTypeId`.\n\nUse MCP prompt `cms_edit_create_article` for guided creation.\n\n## Common sequences\n\n### List all pages\n\n```\n[\"list\", \"--type\", \"page\"]\n```\n\n### Refresh local index (assets, media, templates, taxonomy)\n\nRun before asset filename search, media-by-filename list, or fast template/taxonomy lookup:\n\n```\n[\"index\", \"sync\"]\n[\"index\", \"status\"]\n[\"list\", \"--type\", \"template\"]\n```\n\nAsset search and media-by-filename list **require** the index. Template/taxonomy list prefers the index when built.\n\n### Explore the homepage\n\n```\n[\"open\", \"--page-slug\", \"/\"]\n[\"snapshot\"]\n[\"read\", \"@c1\"]\n```\n\n### Edit copy on a page\n\n```\n[\"open\", \"--page-slug\", \"{{EXAMPLE_SLUG}}\"]\n[\"snapshot\"]\n[\"read\", \"@c1\", \"heading\"]\n[\"set\", \"@c1\", \"heading\", \"Updated heading\"]\n[\"diff\"]\n[\"save\"]\n```\n\n## Field discovery\n\n```\n[\"schema\", \"component\"]\n[\"list\", \"--type\", \"page\"]\n[\"search\", \"pricing\"]\n```\n\n## Grok Build setup\n\nGrok connects to the same hosted MCP URL as Claude. Add to `~/.grok/config.toml`:\n\n```toml\n[mcp_servers.cms-edit-brightline]\nurl = \"https://brightline.content.se.studio/api/mcp\"\nenabled = true\n```\n\nReplace the server name and host with your site's deployment. The URL **must** end with `/api/mcp` (not the site root).\n\n**Connect:**\n\n1. In Grok, run `/mcps` and select your cms-edit server\n2. Press **`i`** to initiate OAuth (not just enable the toggle)\n3. Sign in with Contentful when the browser opens\n4. Verify with `grok mcp doctor <server-name>` or call `cms_edit_version`\n\nOAuth is two steps: Grok obtains an MCP token, then you sign in with Contentful to link your account. If the browser flashes and closes immediately, reconnect after upgrading cms-edit (loopback redirect handling).\n\n## Hosted connector rules\n\n{{HOSTED_RULES}}\n\n## Session tips\n\n- After `open`, subsequent MCP commands share session `_mcp` automatically (stored in Redis on hosted).\n- Use `--session <name>` only when juggling multiple pages — pass it **before** the subcommand on every call: `[\"--session\", \"my-work\", \"open\", \"--page-slug\", \"/\"]`.\n- Snapshot JSON includes `sessionName` and `sessionKey` for debugging.\n- If you see \"No active session\", check `/api/health` (Redis must be ok), then run `open` again. Set `LOG_CMS_SESSION=1` on Vercel and search runtime logs for `[cms-edit:session]` (`redis save ok` / `redis load hit` / `requireSession miss`).\n\n### peek vs open vs read\n\n| Goal | MCP args |\n|------|----------|\n| Edit workflow (session) | `[\"open\", \"--page-slug\", \"/\"]` then `[\"read\", \"@c0\"]` |\n| Preview tree + fields, no session | `[\"peek\", \"--page-slug\", \"/pricing\"]` or `[\"peek\", \"--id\", \"<entry-id>\"]` (JSON includes `rootFields`) |\n| Read fields by entry ID, no session | `[\"read\", \"<entry-id>\", \"--id\"]` or `[\"read\", \"<entry-id>\", \"--id\", \"heading\"]` |\n| Debug session persistence | `[\"diagnose\"]` — reports store, userId, active session key |\n\n- **`peek` is one-shot** — prints tree and **root field values** (`rootFields` in JSON). No session is created; do not chain `read @ref` afterward.\n- **`open` / `nav open` persist** the session. `preview urls /slug` follows that slug even if leftover `_mcp` is another page (JSON may include `sessionRootIgnored`).\n- **Use flat lookup flags** for `open` and `peek` — same as CLI (`--page-slug`, `--id`, etc.). See `help lookup`. `nav open` uses `--nav-id` or `--nav-name` only.\n- **`read @root`** only works after `open` (uses the active session). The hint in peek output is for terminal users who already ran `open`.\n- **Do not invent flags** — there is no `--read`, `--then`, or chained peek+read syntax. One command per `cms_edit` call.\n\n## Troubleshooting\n\n| Problem | What to do |\n|---------|------------|\n| **401 / connection failed** | Reconnect: Claude → Settings → Integrations, or Grok `/mcps` → press `i`. Complete Contentful sign-in. If Grok says “Failed to get tools after auth” or `auth required`, check Vercel `[cms-edit:oauth] auth_failure`. `invalid_refresh_token` is a rotated/missing MCP refresh UUID (fixed in **3.6.1+** by not rotating). `missing_contentful_token` / `invalid_contentful_token` means the CMA cache is gone — **3.5.3+** returns `invalid_token` so Grok re-runs Contentful sign-in. If it still loops, delete that server’s entry from `~/.grok/mcp_credentials.json` and press `i` again. Set `LOG_CMS_OAUTH=1` for connect success traces. |\n| **No cms_edit tool** | Confirm the integration is connected under Settings → Integrations |\n| **No active session** | Do not pass `--space` on hosted. Check `/api/health` — `redis` must be ok. MCP uses `_mcp` session automatically. Manual session: `[\"--session\", \"name\", \"open\", \"--page-slug\", \"/\"]`. Fallback: `[\"read\", \"<entry-id>\", \"--id\"]` |\n| **Wrong space** | Your Contentful account must have access to this deployment's space |\n| **Check version** | Call `cms_edit_version` or `curl /api/health` and read `cmsEdit` |\n\n## Articles\n\n**Recent articles** (by publication `date`, not last-updated):\n\n```json\n[\"list\", \"--type\", \"article\", \"--sort\", \"date\", \"-n\", \"10\"]\n```\n\nAlternative: `[\"search\", \"<keyword>\", \"--type\", \"article\"]` (also sorts by date).\n\n**Draft articles** — `peek` and `open` load drafts via CMA (not published-only). To discover drafts:\n\n```json\n[\"list\", \"--type\", \"article\", \"--sort\", \"date\"]\n[\"search\", \"<title>\", \"--type\", \"article\"]\n[\"list\", \"--type\", \"article\", \"--slug\", \"<slug>\"]\n```\n\nThen open: `[\"peek\", \"--article-slug\", \"<slug>\"]` or `[\"open\", \"--id\", \"<entry-id>\"]`. Use `--article-slug`, not `--page-slug`.\n\n**Published-only** is opt-in: `list --published` and `resolve --published` use the Delivery API index and omit never-published drafts.\n\nAfter upgrading cms-edit, run `[\"index\", \"sync\"]` once so the local index includes article `date` values (schema v4).\n\n**Index freshness on hosted MCP:** The catalog index is shared via Redis (gzipped SQLite blob). `index status` on a cold instance may show **not built** until `list`, `open`, or another index command runs — that is normal. Committed writes (`save`, `create page`, `asset upload`, …) mark the index dirty; the next discovery command refreshes it. In-session `set` edits do not trigger a sync.\n\n## Command reference (cms_edit args)\n\n**Discovery:** `list --type <type>`, `search <query>`, `schema <contentType>`, `index sync`, `index status`\n\n**Reading:** `open --page-slug <slug>`, `open --id <id>`, `peek --page-slug <slug>`, `peek --id <id>`, `snapshot`, `read <ref>`, `read <ref> <field>`, `read <id> --id [field]`\n\n**Diagnostics:** `diagnose` — session store, userId, active session (after open/read issues)\n\n**Editing:** `set <ref> <field> <value>`, `rtf <ref> <field> --content \"<md>\"`, `diff`, `save`, `discard`\n\n**Creating:** `create from-json --json … --dry-run --strict`, `export from-json`, `apply from-json --json …`, `create page` / `create article` for empty shells, `add` for tiny inserts\n\n**Other:** `health`, `version`, `guide [topic]`\n";
+export const HOSTED_GUIDE_TEMPLATE = "# cms-edit MCP guide\n\nUse this guide when editing Contentful draft content through a hosted MCP connector (Claude Desktop Integrations + OAuth, or Grok Build).\n\n{{SITE_INTRO}}\n\n## Safety rules\n\ncms-edit **cannot and will not**:\n\n- Publish or unpublish entries\n- Archive or delete published entries\n\nAll `save` operations create **drafts only**. A human must review and publish in Contentful.\n\n**Assets:** cms-edit **can upload** draft assets via **staged upload** (hosted) or `--url` (public HTTPS last resort). **`--base64` is not supported on hosted MCP.** Always **search and reuse existing assets first** (`index sync`, then `asset search` or `list --type media`). Use `--if-exists-by-filename` when uploading to avoid duplicates.\n\n## MCP tools\n\n| Tool | Purpose |\n|------|---------|\n| `cms_edit` | Run any cms-edit command via an `args` array |\n| `cms_edit_request_staged_upload` | Presigned URL for binaries. Raster/video: then `cms_edit_media` prepare. SVG/PDF/Lottie: then `asset upload --staged` |\n| `cms_edit_version` | Check the installed cms-edit version (no session required) |\n| `cms_edit_help` | Guide topics (`overview`, `workflow`, `commands`, `troubleshooting`, `hosted`) **or** playbooks (`create-from-json`, `fields-html-component`, `fields-component`, …) |\n\n## Customer editor pack (read first)\n\nWhen a site editor pack is deployed, these MCP resources provide **site-specific** context. Read them before creating or editing content:\n\n| Resource | Purpose |\n|----------|---------|\n| `cms-edit://customer/index` | List of all customer resources |\n| `cms-edit://customer/tasks-index` | Intent → task playbook map (filtered for this site) |\n| `cms-edit://customer/capabilities` | Machine-readable list of what this site supports |\n| `cms-edit://customer/checklist` | First-session onboarding steps |\n| `cms-edit://customer/overview` | Site name, space, staging URL, handoff rules |\n| `cms-edit://customer/routing` | Article/tag URL rules for this site |\n| `cms-edit://customer/production-site` | Production URL and how to fetch page markdown (do not guess `.md` URLs) |\n| `cms-edit://customer/components-index` | Component types and when to use each |\n| `cms-edit://customer/collections-index` | Collection types |\n| `cms-edit://customer/brand` | Brand voice and terminology (if configured) |\n| `cms-edit://customer/articles` | Article/publication playbook (if configured) |\n| `cms-edit://customer/defaults` | Featured image rules and article types (if configured) |\n| `cms-edit://customer/site-facts` | Optional canonical business facts (if configured) |\n| `cms-edit://customer/task-*` | Capability playbooks (e.g. `task-media-reuse-and-upload`) |\n\n**Workflow:** read `tasks-index` → `capabilities` → `checklist` → `overview` → `routing` → `production-site` → pick a `task-*` playbook for your goal → `components-index` → then use `cms_edit` commands.\n\n## MCP prompts\n\n| Prompt | Purpose |\n|--------|---------|\n| `cms_edit_verify` | Check version and connectivity |\n| `cms_edit_explore` | List site content and open a page |\n| `cms_edit_edit_page` | Edit a page through diff → save |\n| `cms_edit_edit_article` | Edit an article through diff → save |\n| `cms_edit_index_sync` | Build/refresh local index (assets, media, templates, taxonomy) |\n| `cms_edit_create_page` | Create a page with template selection |\n| `cms_edit_create_article` | Create an article (requires article type; featured image rules) |\n| `cms_edit_create_from_doc` | Create a page from a freeform document (infer → confirm → dry-run) |\n| `cms_edit_create_article_from_doc` | Create an article from a freeform document |\n| `cms_edit_taxonomy` | Import taxonomy entries (`create … --if-not-exists`) |\n| `cms_edit_bulk_seo` | Update SEO meta descriptions (read `task-seo-descriptions`) |\n| `cms_edit_bulk_find_replace` | Bulk find-replace in rich text across entries |\n| `cms_edit_bulk_article_tags` | Update tags on one or more articles |\n| `cms_edit_import_authors` | Create person/author entries from a list or sheet |\n| `cms_edit_media` | Find existing assets or upload new ones (reuse-first) |\n\nRead the full guide anytime via the `cms-edit://guide` resource or `cms_edit_help`.\n\n## End-to-end QA prompt\n\nFor release QA after cms-edit or editor-pack upgrades, use a **copy-paste Claude Desktop test script** (phases: connector → customer resources → discovery → session → schema → media → dry-run writes → negatives). See [e2e-test-prompt.md](./e2e-test-prompt.md) to author one per customer site. Example: Brightline `docs/cms-edit/MCP-E2E-TEST.md`.\n\n## Core workflow\n\nStandard editing flow: **open → snapshot → read → set/rtf → diff → save**\n\nEach step is a separate `cms_edit` call. Sessions persist in Redis between calls (hosted) or on disk (local). MCP auto-uses session `_mcp`.\n\n### 1. Open a page\n\n```\n[\"open\", \"--page-slug\", \"/\"]\n[\"open\", \"--page-slug\", \"/pricing\"]\n[\"open\", \"--id\", \"<entry-id>\"]\n[\"open\", \"--include-chrome\", \"--page-slug\", \"/pricing\"]\n```\n\nDefault `open` / `peek` fetch the page or article tree and the template entry, but **not** template menu/footer/navigation. Edit nav with `[\"nav\", \"open\", \"--nav-id\", \"<id>\"]` or `--nav-name` (not `--id`). Use `--include-chrome` only when you need those trees. Template `menu` is site-wide; `page.menu` overrides one page. Nested nav items are often shared — ask global vs page-specific before `set`.\n\n**Homepage (all SE Studio sites):** Contentful stores the marketing homepage slug as `index`. The live site URL is `/`. In cms-edit, `--page-slug /` and `--page-slug /index` are equivalent (both resolve to `index`). Read `cms-edit://customer/routing` for site-specific exceptions (e.g. Pedestal uses `home`).\n\n### 2. Snapshot — see the content tree\n\n```\n[\"snapshot\"]\n```\n\nOutput shows `@c0`, `@c1`, … refs for each component. Use these refs in `read`, `set`, and `rtf`.\n\n### 3. Read fields\n\n```\n[\"read\", \"@c1\"]\n[\"read\", \"@c1\", \"heading\"]\n```\n\n### 4. Edit fields\n\n```\n[\"set\", \"@c1\", \"heading\", \"New heading text\"]\n[\"rtf\", \"@c1\", \"body\", \"--content\", \"# Title\\n\\nParagraph text.\"]\n```\n\nFor rich text, prefer `--content` with inline markdown. Use `--base64` when special characters cause issues.\n\n### 5. Review and save\n\n```\n[\"diff\"]\n[\"save\"]\n```\n\nAlways run `diff` before `save`. To abandon changes: `[\"discard\"]`.\n\n## When to use Contentful UI\n\n| Task | Where |\n|------|-------|\n| Edit copy, add sections, create pages/articles | **Agent + cms-edit** (drafts) |\n| Upload new images and files | **Agent + cms-edit** — search/reuse first; raster/video via `cms_edit_media` import |\n| Publish approved drafts | **Contentful UI** |\n| Edit copy directly in Contentful | **Avoid** — use cms-edit instead |\n\ncms-edit saves **drafts only**. A colleague must **publish in Contentful** for changes to go live.\n\n## Asset upload (hosted MCP)\n\nHosted connectors have **no server filesystem** and **reject `asset upload --base64`**. Raster/video: staged intake then **media pipeline**. SVG/PDF/Lottie: staged then `asset upload --staged`.\n\n| Source | When | Example |\n|--------|------|---------|\n| **Staged + pipeline** | jpeg/png/webp/gif/mp4 | `cms_edit_media` prepare `stagedUploadIds` → import (required for video) |\n| **`--url`** | Pipeline-ready **images** only, or SVG/PDF | `[\"asset\", \"upload\", \"--url\", \"https://….blob.vercel-storage.com/ready.jpg\"]` |\n| **Staged consume** | SVG/PDF/Lottie | `asset upload --staged` |\n\n**After upload:** note the returned asset `id`, attach to the entry (`asset set @root media <id>` for person portraits), then `diff` → `save`. Upload alone does not update entries.\n\n**PNG** when transparency matters.\n\nFull workflow: `cms-edit://customer/task-media-reuse-and-upload`.\n\n### Common pitfalls (agents)\n\n| Issue | Do this |\n|-------|---------|\n| Singular Media on a component (`visual`) | `set @cN visual <mediaId> --link` — **not** `add … --target visual` |\n| Shared Drive folder has extra files | Import with `sourceNames` or `onlyWithOverrides` after catalog |\n| Prepare says no media / empty folder | Share folder **Viewer** with `media-pipeline-reader@se-media-pipeline.iam.gserviceaccount.com`, then re-prepare once |\n| MCP tool times out ~180s | `read` / `diff` / re-`catalog` before retrying the same write — server may already have saved |\n| `preview urls` has no Live Preview URL | Use staging + classic entry editor; row includes `unavailableReason` when CMA preview fails |\n| Reuse-first import matches old fileNames | dry-run first; set `overrides[].fileName` for unique names when duplication is required |\n\n### Staged upload\n\nKeeps binary **out of MCP tool args**. The agent uploads via `curl`, then passes only a short `uploadId`.\n\n1. `cms_edit_request_staged_upload` with `fileName`, `mimeType`, optional `byteLength`, optional `title`\n2. Run the returned `curlCommand` in a shell (POST multipart to `uploadUrl`)\n3. Follow `consumeArgs`: **media prepare** (`cms_edit_media` / `media prepare --staged-ids`) for jpeg/png/webp/gif/mp4; **asset upload --staged** for SVG/PDF/Lottie\n4. For pipeline jobs: wait → catalog → **import** (required for video). Then link and save\n\nHosted limit: **4 MB** per staged file (Vercel body limit). Rate limit: **1000 upload URLs per user per hour**. TTL: **60 minutes**. Raster/video staged files are intake (GET by the pipeline); they are not consumed into CMA.\n\n## Featured images (articles)\n\nRead `cms-edit://customer/defaults` when present. Otherwise:\n\n1. **Custom** — set `featuredImage` or `featuredImageAssetFilename` on the article when the editor provides a specific image\n2. **Article type** — if omitted, cms-edit inherits the linked **articleType** `featuredImage`\n3. **Site default** — from `editor.defaults` in the site config (also in `defaults.json` in the editor pack)\n\nRun `[\"index\", \"sync\"]` before resolving assets by filename.\n\n## JSON payloads (filesystem vs no filesystem)\n\nHosted MCP cannot see your disk. Same commands either way:\n\n- **Filesystem client:** write `page.json`, then `cms_edit [\"create\", \"from-json\", \"--json-base64\", \"<base64 of file>\", \"--dry-run\", \"--strict\"]`\n- **No filesystem (Desktop):** put the JSON in `--json` on the tool call\n\nNever use a local `cms-edit` binary, `--file`, or stdin for content.\n\n## Create content\n\n### Create a page\n\n```\n[\"schema\", \"component\"]\n[\"help\", \"create-from-json\"]\n[\"create\", \"from-json\", \"--json\", \"{...}\", \"--dry-run\", \"--strict\"]\n[\"create\", \"from-json\", \"--json\", \"{...}\"]\n```\n\nUse `\"components\"` (not `\"content\"`) and nested `\"type\"` (not `\"componentType\"`). Peek a live entry or `schema <ct>` before authoring field ids.\n\nEmpty `components` fails validation. Incremental `add`/`set` is for tiny patches only.\n\n### Create an article\n\nEvery article requires an **articleType**:\n\n```\n[\"index\", \"sync\"]\n[\"list\", \"--type\", \"articleType\"]\n[\"create\", \"article\", \"--slug\", \"resources/blog/my-post\", \"--title\", \"My Post\", \"--article-type-slug\", \"resources/blog\"]\n```\n\nOr `create from-json` with `\"type\": \"article\"` and `articleTypeSlug` / `articleTypeId`.\n\nUse MCP prompt `cms_edit_create_article` for guided creation.\n\n## Common sequences\n\n### List all pages\n\n```\n[\"list\", \"--type\", \"page\"]\n```\n\n### Refresh local index (assets, media, templates, taxonomy)\n\nRun before asset filename search, media-by-filename list, or fast template/taxonomy lookup:\n\n```\n[\"index\", \"sync\"]\n[\"index\", \"status\"]\n[\"list\", \"--type\", \"template\"]\n```\n\nAsset search and media-by-filename list **require** the index. Template/taxonomy list prefers the index when built.\n\n### Explore the homepage\n\n```\n[\"open\", \"--page-slug\", \"/\"]\n[\"snapshot\"]\n[\"read\", \"@c1\"]\n```\n\n### Edit copy on a page\n\n```\n[\"open\", \"--page-slug\", \"{{EXAMPLE_SLUG}}\"]\n[\"snapshot\"]\n[\"read\", \"@c1\", \"heading\"]\n[\"set\", \"@c1\", \"heading\", \"Updated heading\"]\n[\"diff\"]\n[\"save\"]\n```\n\n## Field discovery\n\n```\n[\"schema\", \"component\"]\n[\"list\", \"--type\", \"page\"]\n[\"search\", \"pricing\"]\n```\n\n## Grok Build setup\n\nGrok connects to the same hosted MCP URL as Claude. Add to `~/.grok/config.toml`:\n\n```toml\n[mcp_servers.cms-edit-brightline]\nurl = \"https://brightline.content.se.studio/api/mcp\"\nenabled = true\n```\n\nReplace the server name and host with your site's deployment. The URL **must** end with `/api/mcp` (not the site root).\n\n**Connect:**\n\n1. In Grok, run `/mcps` and select your cms-edit server\n2. Press **`i`** to initiate OAuth (not just enable the toggle)\n3. Sign in with Contentful when the browser opens\n4. Verify with `grok mcp doctor <server-name>` or call `cms_edit_version`\n\nOAuth is two steps: Grok obtains an MCP token, then you sign in with Contentful to link your account. If the browser flashes and closes immediately, reconnect after upgrading cms-edit (loopback redirect handling).\n\n## Hosted connector rules\n\n{{HOSTED_RULES}}\n\n## Session tips\n\n- After `open`, subsequent MCP commands share session `_mcp` automatically (stored in Redis on hosted).\n- Use `--session <name>` only when juggling multiple pages — pass it **before** the subcommand on every call: `[\"--session\", \"my-work\", \"open\", \"--page-slug\", \"/\"]`.\n- Snapshot JSON includes `sessionName` and `sessionKey` for debugging.\n- If you see \"No active session\", check `/api/health` (Redis must be ok), then run `open` again. Set `LOG_CMS_SESSION=1` on Vercel and search runtime logs for `[cms-edit:session]` (`redis save ok` / `redis load hit` / `requireSession miss`).\n\n### peek vs open vs read\n\n| Goal | MCP args |\n|------|----------|\n| Edit workflow (session) | `[\"open\", \"--page-slug\", \"/\"]` then `[\"read\", \"@c0\"]` |\n| Preview tree + fields, no session | `[\"peek\", \"--page-slug\", \"/pricing\"]` or `[\"peek\", \"--id\", \"<entry-id>\"]` (JSON includes `rootFields`) |\n| Read fields by entry ID, no session | `[\"read\", \"<entry-id>\", \"--id\"]` or `[\"read\", \"<entry-id>\", \"--id\", \"heading\"]` |\n| Debug session persistence | `[\"diagnose\"]` — reports store, userId, active session key |\n\n- **`peek` is one-shot** — prints tree and **root field values** (`rootFields` in JSON). No session is created; do not chain `read @ref` afterward.\n- **`open` / `nav open` persist** the session. `preview urls /slug` follows that slug even if leftover `_mcp` is another page (JSON may include `sessionRootIgnored`).\n- **Use flat lookup flags** for `open` and `peek` — same as CLI (`--page-slug`, `--id`, etc.). See `help lookup`. `nav open` uses `--nav-id` or `--nav-name` only.\n- **`read @root`** only works after `open` (uses the active session). The hint in peek output is for terminal users who already ran `open`.\n- **Do not invent flags** — there is no `--read`, `--then`, or chained peek+read syntax. One command per `cms_edit` call.\n\n## Troubleshooting\n\n| Problem | What to do |\n|---------|------------|\n| **401 / connection failed** | Reconnect: Claude → Settings → Integrations, or Grok `/mcps` → press `i`. Complete Contentful sign-in. If Grok says “Failed to get tools after auth” or `auth required`, check Vercel `[cms-edit:oauth] auth_failure`. `invalid_refresh_token` is a rotated/missing MCP refresh UUID (fixed in **3.6.1+** by not rotating). `missing_contentful_token` / `invalid_contentful_token` means the CMA cache is gone — **3.5.3+** returns `invalid_token` so Grok re-runs Contentful sign-in. If it still loops, delete that server’s entry from `~/.grok/mcp_credentials.json` and press `i` again. Set `LOG_CMS_OAUTH=1` for connect success traces. |\n| **No cms_edit tool** | Confirm the integration is connected under Settings → Integrations |\n| **No active session** | Do not pass `--space` on hosted. Check `/api/health` — `redis` must be ok. MCP uses `_mcp` session automatically. Manual session: `[\"--session\", \"name\", \"open\", \"--page-slug\", \"/\"]`. Fallback: `[\"read\", \"<entry-id>\", \"--id\"]` |\n| **Wrong space** | Your Contentful account must have access to this deployment's space |\n| **Check version** | Call `cms_edit_version` or `curl /api/health` and read `cmsEdit` |\n\n## Articles\n\n**Recent articles** (by publication `date`, not last-updated):\n\n```json\n[\"list\", \"--type\", \"article\", \"--sort\", \"date\", \"-n\", \"10\"]\n```\n\nAlternative: `[\"search\", \"<keyword>\", \"--type\", \"article\"]` (also sorts by date).\n\n**Draft articles** — `peek` and `open` load drafts via CMA (not published-only). To discover drafts:\n\n```json\n[\"list\", \"--type\", \"article\", \"--sort\", \"date\"]\n[\"search\", \"<title>\", \"--type\", \"article\"]\n[\"list\", \"--type\", \"article\", \"--slug\", \"<slug>\"]\n```\n\nThen open: `[\"peek\", \"--article-slug\", \"<slug>\"]` or `[\"open\", \"--id\", \"<entry-id>\"]`. Use `--article-slug`, not `--page-slug`.\n\n**Published-only** is opt-in: `list --published` and `resolve --published` use the Delivery API index and omit never-published drafts.\n\nAfter upgrading cms-edit, run `[\"index\", \"sync\"]` once so the local index includes article `date` values (schema v4).\n\n**Index freshness on hosted MCP:** The catalog index is shared via Redis (gzipped SQLite blob). `index status` on a cold instance may show **not built** until `list`, `open`, or another index command runs — that is normal. Committed writes (`save`, `create page`, `asset upload`, …) mark the index dirty; the next discovery command refreshes it. In-session `set` edits do not trigger a sync.\n\n## Command reference (cms_edit args)\n\n**Discovery:** `list --type <type>`, `search <query>`, `schema <contentType>`, `index sync`, `index status`\n\n**Reading:** `open --page-slug <slug>`, `open --id <id>`, `peek --page-slug <slug>`, `peek --id <id>`, `snapshot`, `read <ref>`, `read <ref> <field>`, `read <id> --id [field]`\n\n**Diagnostics:** `diagnose` — session store, userId, active session (after open/read issues)\n\n**Editing:** `set <ref> <field> <value>`, `rtf <ref> <field> --content \"<md>\"`, `diff`, `save`, `discard`\n\n**Creating:** `create from-json --json … --dry-run --strict`, `export from-json`, `apply from-json --json …`, `create page` / `create article` for empty shells, `add` for tiny inserts\n\n**Other:** `health`, `version`, `guide [topic]`\n";
 //# sourceMappingURL=hosted-guide.js.map

package/dist/generated/hosted-guide.js.map

@@ -1 +1 @@
-{"version":3,"file":"hosted-guide.js","sourceRoot":"","sources":["../../src/generated/hosted-guide.ts"],"names":[],"mappings":"AAAA,4EAA4E;AAC5E,MAAM,CAAC,MAAM,qBAAqB,GAAG,omkBAAomkB,CAAC"}
+{"version":3,"file":"hosted-guide.js","sourceRoot":"","sources":["../../src/generated/hosted-guide.ts"],"names":[],"mappings":"AAAA,4EAA4E;AAC5E,MAAM,CAAC,MAAM,qBAAqB,GAAG,6nkBAA6nkB,CAAC"}

package/dist/generated/version.d.ts

@@ -1,3 +1,3 @@
 export declare const CMS_EDIT_PACKAGE_NAME = "@se-studio/contentful-cms";
-export declare const CMS_EDIT_VERSION = "3.10.1";
+export declare const CMS_EDIT_VERSION = "3.10.2";
 //# sourceMappingURL=version.d.ts.map

package/dist/generated/version.js

@@ -1,4 +1,4 @@
 // Generated by scripts/generate-embedded-assets.mjs — do not edit manually.
 export const CMS_EDIT_PACKAGE_NAME = "@se-studio/contentful-cms";
-export const CMS_EDIT_VERSION = "3.10.1";
+export const CMS_EDIT_VERSION = "3.10.2";
 //# sourceMappingURL=version.js.map

package/docs/editor-tasks/create-from-document.md

@@ -16,7 +16,7 @@
 
 - Fetch package materials (Google Drive MCP, paste, Figma link, zip, production ref, **or designer HTML**)
 - Read `cms-edit://customer/components-index`, `routing`, and for articles **`articles`**
-- **Designer HTML / interactive prototype:** if `cms-edit://customer/task-create-html-component` (or another site HtmlComponent playbook) exists, **switch to that playbook** and stop this one. Otherwise follow the HtmlComponent fork in Phase 2 — do not flatten custom CSS/JS/scroll into registered cards.
+- **Designer HTML / interactive prototype / vendor JS:** switch to `cms-edit://customer/task-create-html-component` and stop this one. Do not flatten custom CSS/JS/scroll into registered cards. Libraries go on `externalScripts`, never `<script src>` or `createElement('script')`.
 - **Non-trivial packages:** complete **`task-source-readiness-review`** first (inventory → beat sheet → verdict → light human yes)
 
 ## Steps
@@ -44,7 +44,7 @@
    - If the source is custom HTML/JS, sticky/scroll, diagrams, or a private layout shell, and this space has `htmlComponent`, map that beat to **`htmlComponent`** — not cards/lists that only match the copy.
    - Mixed pages are expected. Merge consecutive custom sections that share CSS/JS/sticky into **one** HtmlComponent. Do not emit one HtmlComponent per `<section>`.
    - Prefer **source order** over CMS convenience (do not invent side-by-side pairs unless the source is a true pair).
-3. **Propose plan** (can reuse the soft-proof): slug, title, template/article type, ordered component list (registered **and** `htmlComponent` where needed), gap/placeholder list, **asset map** — **ask user to confirm** with a light **yes** if not already confirmed. For HtmlComponent beats: extract `data:` media first; run `schema htmlComponent` for this space’s field IDs (`html` vs `htmlContent`).
+3. **Propose plan** (can reuse the soft-proof): slug, title, template/article type, ordered component list (registered **and** `htmlComponent` where needed), gap/placeholder list, **asset map** — **ask user to confirm** with a light **yes** if not already confirmed. For HtmlComponent beats: extract `data:` media first; run `schema htmlComponent` for this space’s field IDs (`html` vs `htmlContent`); list vendor URLs for `externalScripts` (see `task-create-html-component`).
 4. For pages: `list --type template` — pick template; check template `preContent` / `postContent` so heroes are not duplicated.
 5. For articles: resolve `articleType`; apply site SEO and lab rules from `articles` playbook.
 6. Build `create from-json` payload **or** prefer incremental `open` → `add` / `set` / `rtf` for large multi-block work. Link visuals using the pack map from Phase 1.
@@ -81,4 +81,4 @@ Use MCP prompts `cms_edit_create_from_doc` or `cms_edit_create_article_from_doc`
 - `task-create-page`
 - `task-media-reuse-and-upload`
 - `task-preview-verify`
-- Site HtmlComponent playbook when present (`cms-edit://customer/task-create-html-component`)
+- `cms-edit://customer/task-create-html-component`

package/docs/editor-tasks/create-html-component.md

@@ -0,0 +1,135 @@
+# Capability: Create from designer HTML (HtmlComponent)
+
+## Intent phrases
+
+- build from designer HTML
+- HTML dump
+- create from HTML prototype
+- import HTML landing page
+- HtmlComponent from HTML
+- custom HTML page
+- load HtmlComponent external scripts
+
+## Requires capabilities
+
+`pages` or `articles`
+
+## Prerequisites
+
+- Read `cms-edit://customer/pages` (or `articles`) and `cms-edit://customer/components-index`
+- `cms_edit ["schema", "htmlComponent"]` — use this space’s field IDs (`rawHtml` vs `htmlContent`, `customJs` vs `jsContent`). **`externalScripts` is the same field id on every space.**
+- Designer HTML / prototype available (file, paste, or Drive). **Do not** paste a multi-megabyte dump with `data:` URIs into Contentful.
+- For ordinary Google Doc / Figma-to-registered-stack briefs with no custom CSS/JS, use `task-create-from-document` instead.
+
+## When to use this playbook
+
+Use this when the source is a **designer HTML dump**, interactive prototype, or a block that needs a **third-party JS library**. Do not flatten custom CSS/JS/scroll into registered cards because the copy looks like a card grid.
+
+## Per-section decision
+
+For **each** source section, ask: do **layout and interaction** match a registered type?
+
+| Fit | Action |
+|-----|--------|
+| Strong (same pattern) | Use the registered type. |
+| Medium (copy matches, chrome differs) | Ask the human: site chrome vs keep HtmlComponent. |
+| Weak / none | HtmlComponent. Custom sticky scroll, diagrams, widgets, vendor JS. |
+
+Mock **nav and footer** in the HTML: strip. The page **template** supplies chrome.
+
+## Split HtmlComponents (mixed pages)
+
+Mixed pages are expected: registered types **plus** one or more `htmlComponent` entries.
+
+**Do split** when a section matches a registered type, remaining HTML would exceed field limits, or adjacent custom blocks **do not share** JS/CSS/sticky state.
+
+**Do not split** when one behaviour spans the markup (sticky rail + stacked cards) or several sections share one stylesheet / SVG sprite / library.
+
+Rules:
+
+1. Registered match → registered type.
+2. Merge consecutive custom sections that share CSS/JS into **one** HtmlComponent.
+3. Start a new HtmlComponent only at a clean boundary (no shared script, no sticky parent).
+4. Never one HtmlComponent per `<section>` by default.
+
+## External scripts (non-negotiable)
+
+Put third-party libraries in **`externalScripts`**, not `<script src>` in HTML and not `createElement('script')` inside `customJs`.
+
+Allowed values:
+
+- Same-origin paths under `/html-components/` that end in `.js` (no `..`, no query, no hash). Prefer files in the site repo at `public/html-components/<feature>/….js`.
+- HTTPS URLs whose hostname is already listed on the site’s `CmsRendererConfig.htmlComponentExternalScriptHosts`. If the site sets CSP, that host must also be in `script-src`.
+
+If the host is **not** allowlisted, **stop** — do not invent a CDN load and do not patch renderer config without a human.
+
+Rejected (renderer drops these; preview/dev may warn): `http:`, `javascript:`, protocol-relative `//…`, path `..`, query/hash on same-origin paths.
+
+Max **10** URLs per entry. Duplicate URLs on one page share one `next/script` id. Externals load with `scriptStrategy` (default `afterInteractive`). **`customJs` runs only after every listed src is ready**; if any src fails, inline JS is not injected.
+
+### How to set `externalScripts`
+
+It is an **Array of Symbol**, not a string. Scalar `set` is rejected.
+
+```text
+cms_edit ["set", "@cN", "externalScripts", "--json", "[\"/html-components/feature/lib.js\"]"]
+```
+
+On `create from-json`, put the same array under `fields.externalScripts`. Run `schema htmlComponent` first — unknown keys fail (`html` vs `htmlContent`).
+
+Then set HTML/CSS/JS with `--content` / `--content-base64` on `rawHtml`/`customCss`/`customJs` (or this space’s dual-write ids).
+
+## Media first (mandatory)
+
+1. Extract binary media from `data:` URIs and local files. Upload via `task-media-reuse-and-upload` (search/reuse first).
+2. Rewrite HTML `src` / `poster` to Contentful or CDN URLs. Never leave base64 in the field.
+3. `markdownContent` = user-visible copy and structure. Omit layout chrome and JS behaviour.
+
+## HtmlComponent authoring
+
+1. Follow the site HTML style guide (`docs/cms-guidelines/html-component-style-guide.md` when available).
+2. Scope all JS queries to injected `hcRoot`.
+3. `isHero` — `true` if this block contains the page `<h1>`. `fullWidth` — as needed for full-bleed sections.
+4. Set `cmsLabel` immediately (never leave “New Generic”).
+5. Dual-write legacy + core field IDs when `schema htmlComponent` shows both.
+6. Draft only — never publish unless the human explicitly asks.
+
+## Steps
+
+1. Inventory sections. For each: registered / ask / HtmlComponent. Merge custom blocks that share CSS/JS.
+2. List vendor libraries. Copy same-origin files into `public/html-components/<feature>/` or confirm the HTTPS host is already allowlisted.
+3. Post a soft-proof (ordered stack + `externalScripts` URLs) and get a light human **yes**.
+4. Media pipeline: extract, upload, rewrite URLs.
+5. Create or open the page/article. Add registered components where decided.
+6. For each HtmlComponent: `add htmlComponent` (or `create from-json --dry-run --strict` with this space’s field keys).
+7. Set `externalScripts` with `--json`. Set HTML/CSS/JS. Set `cmsLabel`.
+8. `diff` → `save` (draft).
+9. `cms_edit ["preview", "urls", "<slug>"]` — follow `task-preview-verify`. Confirm libraries load (no `createElement('script')`, no dropped-URL warnings).
+
+## Confirmation gates
+
+1. Soft-proof (mixed stack + script URLs) + human yes
+2. Media extracted (no `data:` in saved HTML)
+3. `schema htmlComponent` field IDs used
+4. Libraries only via `externalScripts`
+5. `diff` before `save`
+
+## Out of scope
+
+- Publish
+- Flattening custom interaction into cards/lists because copy looks similar
+- One HtmlComponent per `<section>` by default
+- Rewriting source copy
+- Pasting the raw designer file into HTML
+- Adding a new CDN host to `htmlComponentExternalScriptHosts` without a human
+- Scalar `set` of a single URL string onto `externalScripts`
+
+## Related resources
+
+- `cms-edit://customer/pages`
+- `cms-edit://customer/task-create-from-document`
+- `cms-edit://customer/task-create-page`
+- `cms-edit://customer/task-media-reuse-and-upload`
+- `cms-edit://customer/task-preview-verify`
+- `cms-edit://customer/task-publish-handoff`
+- `cms_edit ["help", "fields-html-component"]`

package/docs/editor-tasks/source-readiness-review.md

@@ -45,11 +45,11 @@ If title, order, or hero asset **conflict** and the user has not decided → **N
    |---|------|-------------------|-----------------------------------|----------------------|
    | 1 | … | … | … | … |
 
-3. **Build beat sheet** — ordered narrative + visual beats. For each beat: use a **registered** component/collection type when **layout and interaction** match (`components-index` / site playbook). If the source is designer HTML, custom CSS/JS, sticky/scroll, or a diagram that no registered type can reproduce, map to **`htmlComponent`** when that type exists in the space (and follow a site `task-create-html-component` playbook if present). Mixed stacks are valid. Merge consecutive custom HTML beats that share CSS/JS. Prefer **source order** over CMS convenience.
+3. **Build beat sheet** — ordered narrative + visual beats. For each beat: use a **registered** component/collection type when **layout and interaction** match (`components-index` / site playbook). If the source is designer HTML, custom CSS/JS, sticky/scroll, or a diagram that no registered type can reproduce, map to **`htmlComponent`** when that type exists in the space and follow **`task-create-html-component`**. Mixed stacks are valid. Merge consecutive custom HTML beats that share CSS/JS. Prefer **source order** over CMS convenience. Vendor libraries belong on `externalScripts`.
 4. **Score verdict** using the threshold below.
 5. **Post one-screen soft-proof** — numbered final `content` (or page stack) order + gap list + verdict.
 6. **Ask for a light human confirm** — accept **yes** / **go** / **proceed** (or equivalent). Do not build on silence if the package is non-trivial.
-7. On **GO** or **GO-WITH-GAPS** + yes → hand off to `task-create-html-component` (if the site has it and the source is designer HTML), otherwise `task-create-from-document` / `task-create-article` / `task-create-page` with the soft-proof as the plan.
+7. On **GO** or **GO-WITH-GAPS** + yes → hand off to `task-create-html-component` when the source is designer HTML / vendor JS, otherwise `task-create-from-document` / `task-create-article` / `task-create-page` with the soft-proof as the plan.
 8. On **NO-GO** → stop; list what is blocking (do not invent structure to unblock).
 
 ## Gap threshold
@@ -114,4 +114,4 @@ Copy: unchanged from source (list any human decisions only)
 - `task-media-pipeline-prepare`
 - `task-media-reuse-and-upload`
 - `task-preview-verify`
-- Site `task-create-html-component` when present
+- `task-create-html-component`

package/docs/help/create-from-json.md

@@ -17,7 +17,7 @@ cms_edit ["create", "from-json", "--json", "{...}", "--dry-run", "--strict"]
 - Nested tree key is **`components`**, not `content` (`content` is a page field — empty page).
 - Discriminator is **`type`**, not `componentType` / `collectionType`.
 - **`components` must be non-empty**.
-- Unknown nested keys fail (`html` vs `htmlContent` — run `schema htmlComponent` for this space).
+- Unknown nested keys fail (`html` vs `htmlContent` — run `schema htmlComponent` for this space). Put vendor libraries on `fields.externalScripts` (string array), not inside HTML.
 - CTA text is **`label`**, not `linkText`.
 - Quoted `set @c1:field="value"` is rejected (quotes would be stored). Use `["set", "@c1", "field", "value"]`.
 
@@ -403,3 +403,29 @@ Once created:
 3. Publish in the Contentful web app when ready.
 
 The tool prints the Contentful web app URL for the new entry after creation.
+
+---
+
+## HtmlComponent example
+
+Run `schema htmlComponent` first. Field ids below are the core names; some spaces use `htmlContent` / `cssContent` / `jsContent` instead (or dual-write both). Vendor libraries belong on `externalScripts`, not inside HTML.
+
+```json
+{
+  "type": "page",
+  "slug": "/lab/html-widget",
+  "title": "HTML widget",
+  "components": [
+    {
+      "contentType": "htmlComponent",
+      "cmsLabel": "Widget — jsPDF demo",
+      "fields": {
+        "rawHtml": "<div class=\"widget\"><button type=\"button\" class=\"export\">Export</button></div>",
+        "customJs": "hcRoot.querySelector('.export').addEventListener('click', function () { window.jspdf.jsPDF(); });",
+        "externalScripts": ["/html-components/widget/jspdf.umd.min.js"],
+        "markdownContent": "Export widget"
+      }
+    }
+  ]
+}
+```

package/docs/help/fields-html-component.md

@@ -0,0 +1,78 @@
+# cms-edit — HtmlComponent Field Reference
+
+An `htmlComponent` entry is developer-authored HTML, optional scoped CSS, optional inline JS, and optional **external script URLs**. It is not a registered `component` variant.
+
+Run `cms-edit schema htmlComponent` for this space’s field IDs. Some spaces dual-write legacy names (`htmlContent` / `cssContent` / `jsContent`) plus core ids (`rawHtml` / `customCss` / `customJs`). **`externalScripts` is the same id on every space.** Unknown JSON keys fail (`html` vs `htmlContent`).
+
+See `cms-edit://customer/task-create-html-component` for the full create playbook.
+
+---
+
+## Fields
+
+| Field | Type | Required | Description |
+|---|---|---|---|
+| `cmsLabel` | Symbol | Yes | Internal label. Set immediately. |
+| `rawHtml` / `htmlContent` | Text | Yes | Markup. No `<script src>`. |
+| `customCss` / `cssContent` | Text | No | Auto-scoped to `[data-hc-id]`. |
+| `customJs` / `jsContent` | Text | No | Inline JS via `next/script`. Scope queries to injected `hcRoot`. |
+| `externalScripts` | Array of Symbol | No | Allowlisted script URLs loaded **before** `customJs`. Max 10. |
+| `markdownContent` | Text | No | Plain text for search / markdown export. |
+| `scriptStrategy` | Symbol | No | `afterInteractive` (default), `lazyOnload`, `beforeInteractive`. Applies to externals and inline JS. |
+| `isHero` | Boolean | No | `true` when the HTML contains the page `<h1>`. |
+| `fullWidth` | Boolean | No | Viewport-breakout layout when the site renderer supports it. |
+| `excludeFromSearch` | Boolean | No | Skip search even if `markdownContent` is set. |
+| `loadScriptOnce` | Boolean | No | Dedupe inline script id per entry (default true). External URLs already share one `next/script` id per URL. |
+
+---
+
+## `externalScripts` allowlist
+
+Always allowed: same-origin paths matching `/html-components/` + a `.js` filename (no `..`, no query, no hash). Prefer files in the site repo at `public/html-components/<feature>/….js`.
+
+HTTPS URLs are allowed only when the hostname is in `CmsRendererConfig.htmlComponentExternalScriptHosts`. If the site has CSP, that host must also be in `script-src`. If the host is not allowlisted, **stop** — do not invent a CDN load.
+
+Rejected: `http:`, `javascript:`, `//…`, path traversal, query/hash on same-origin paths. Disallowed URLs are not loaded (preview/dev may warn). Duplicates are dropped at render. `customJs` waits until every listed src is ready; a failed src skips inline JS.
+
+Do **not** put libraries in HTML `<script>` tags or `createElement('script')` inside `customJs`.
+
+---
+
+## Setting fields
+
+HTML/CSS/JS (long text):
+
+```
+cms_edit ["set", "@cN", "rawHtml", "--content-base64", "<b64>"]
+```
+
+`externalScripts` is an array — **not** a scalar string. Scalar `set` is rejected.
+
+```
+cms_edit ["set", "@cN", "externalScripts", "--json", "[\"/html-components/feature/lib.js\"]"]
+```
+
+On `create from-json`:
+
+```json
+{
+  "contentType": "htmlComponent",
+  "cmsLabel": "Pricing grid",
+  "fields": {
+    "rawHtml": "<div class=\"hc-root\">…</div>",
+    "customJs": "hcRoot.querySelector('.buy')",
+    "externalScripts": ["/html-components/pricing-grid/lib.js"],
+    "markdownContent": "Pricing grid"
+  }
+}
+```
+
+Use this space’s field ids from `schema htmlComponent` (`rawHtml` vs `htmlContent`). Dual-write both when the schema shows both.
+
+---
+
+## Related
+
+- `cms-edit://customer/task-create-html-component`
+- `cms-edit help create-from-json`
+- `cms-edit schema htmlComponent`

package/docs/help/fields.md

@@ -10,7 +10,7 @@ This page lists the content types managed by `cms-edit` and links to their field
 |---|---|---|
 | `page` | Web pages with a slug and content tree | `cms-edit help fields-page` |
 | `article` | Blog posts and news items | `cms-edit help fields-article` |
-| `component` | Individual content sections (hero, text block, card, etc.) | `cms-edit help fields-component` |
+| `htmlComponent` | Developer-authored HTML / CSS / JS blocks | `cms-edit help fields-html-component` |
 | `collection` | Groups of components (card grids, timelines, etc.) | `cms-edit help fields-collection` |
 | `link` | CTA button entries | Managed via `cms-edit links` |
 | `template` | Reusable page shell (header, footer slots) | Used via `templateId` in `create from-json` |
@@ -57,4 +57,5 @@ Run `cms-edit snapshot` to see current refs.
 - `cms-edit help fields-page` — page entry fields
 - `cms-edit help fields-article` — article entry fields
 - `cms-edit help fields-component` — component entry fields
+- `cms-edit help fields-html-component` — HtmlComponent (raw HTML, `externalScripts`)
 - `cms-edit help fields-collection` — collection entry fields

package/docs/help/index.md

@@ -29,6 +29,7 @@ Available topics for `cms-edit help <topic>`:
 | `fields-page` | All fields on a `page` entry |
 | `fields-article` | All fields on an `article` entry |
 | `fields-component` | All fields on a `component` entry |
+| `fields-html-component` | HtmlComponent fields including `externalScripts` |
 | `fields-collection` | All fields on a `collection` entry |
 | `fields-taxonomy` | All fields on `tagType`, `articleType`, and `tag` entries |
 | `index-sync` | Build and use the local content index (CDA sync) |

package/docs/mcp/hosted-guide.md

@@ -22,7 +22,7 @@ All `save` operations create **drafts only**. A human must review and publish in
 | `cms_edit` | Run any cms-edit command via an `args` array |
 | `cms_edit_request_staged_upload` | Presigned URL for binaries. Raster/video: then `cms_edit_media` prepare. SVG/PDF/Lottie: then `asset upload --staged` |
 | `cms_edit_version` | Check the installed cms-edit version (no session required) |
-| `cms_edit_help` | Guide topics (`overview`, `workflow`, `commands`, `troubleshooting`, `hosted`) **or** playbooks (`create-from-json`, `fields-component`, …) |
+| `cms_edit_help` | Guide topics (`overview`, `workflow`, `commands`, `troubleshooting`, `hosted`) **or** playbooks (`create-from-json`, `fields-html-component`, `fields-component`, …) |
 
 ## Customer editor pack (read first)
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@se-studio/contentful-cms",
-  "version": "3.10.1",
+  "version": "3.10.2",
   "description": "CLI tool for AI agents to read and edit Contentful draft content without publish permissions",
   "repository": {
     "type": "git",