@collective-intelligence/external-apps-server-sdk 0.3.8 → 0.3.9

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/AGENTS.md CHANGED
@@ -49,3 +49,4 @@ Primitives, the Revalidate Route and the Effect services. The package
49
49
  ## Platform guide
50
50
 
51
51
  - [docs/guide/index.md](docs/guide/index.md): read before choosing CI primitives for a solution or demo, or when a task needs the platform model (architecture, access, data, AI, template).
52
+ - [docs/guide/platform/built-ins.md](docs/guide/platform/built-ins.md): read before declaring a tool, workflow node, trigger, connector, connection, schema or notification channel, to reuse the `builtin.` FQN that CI already ships and to check which built-in packages an Organization has installed.
@@ -67,7 +67,7 @@ declareTool({
67
67
 
68
68
  `usageScopes` decides where the tool can appear: in a Skill, in chat, or in automation. The template needs `automation` so that the Lead qualifier Agent can use the tool. When you omit `usageScopes`, the tool appears in all three places, and an explicit empty list makes it unavailable everywhere. The model reads `toolGuidance.useWhen` and `toolGuidance.inputRequirementsSummary` to decide when to call the tool and what to pass, so write both for the model.
69
69
 
70
- A package can also bind built-in tools for web search (Exa or Tavily, which need an API key), website scraping, image generation, email, notifications, OCR, speech transcription and synthesis, and report building. External applications can use `publish-run-event`.
70
+ A package can also bind built-in tools for web search (Exa or Tavily, which need an API key), website scraping, image generation, email, notifications, OCR, speech transcription and synthesis, and report building. External applications can use `publish-run-event`. The [Built-in catalog](../platform/built-ins.md) lists every built-in tool and says which ones are ready after install.
71
71
 
72
72
  ## Skills
73
73
 
@@ -2,7 +2,7 @@
2
2
 
3
3
  Connect outside systems, map their fields to Knowledge schemas, and run and watch imports.
4
4
 
5
- A connector keeps a Data resource in sync with an outside system. CI has two kinds. A Knowledge connector imports records from a source such as Notion, HubSpot or a SQL database into Knowledge records. A Documents connector imports files from Microsoft 365 or Google Drive into one Connected Collection of Documents. The [Data resources](index.md) overview compares Knowledge, Datasets and Documents, and the [Documents](documents.md) page covers Document Collections.
5
+ A connector keeps a Data resource in sync with an outside system. CI has two kinds. A Knowledge connector imports records from a source such as Notion, HubSpot or a SQL database into Knowledge records. A Documents connector imports files from Microsoft 365 or Google Drive into one Connected Collection of Documents. The [Data resources](index.md) overview compares Knowledge, Datasets and Documents, and the [Documents](documents.md) page covers Document Collections. The [Built-in catalog](../platform/built-ins.md) lists every connector and connection type that CI ships.
6
6
 
7
7
  A Knowledge import passes through several stages: CI reads the source page by page, stages the records, commits the accepted ones into Knowledge, and then indexes them for search. An app or agent that reads imported data has to wait for the stage it needs. Most of this page covers Knowledge connectors. The last section compares them with Documents connectors.
8
8
 
@@ -40,7 +40,7 @@ Executable kinds also take an `implementation` class from `@collective-intellige
40
40
  | `declareEventListener` | Selectors plus `normalizeWebhook` or `poll` |
41
41
  | `declareNotificationChannel` | `deliver` |
42
42
 
43
- A component refers to another component of the same package through the handle that its `declare*` call returned. For a built-in component or one from another package, pass the FQN string instead, for example `"builtin.core.knowledge.tools.query"`. TypeScript does not check these strings. Install resolves them and rolls the package back when one does not exist, and the SDK publishes no list of built-in FQNs, so the template pins its FQNs in a `bun test` that catches a typo before deploy.
43
+ A component refers to another component of the same package through the handle that its `declare*` call returned. For a built-in component or one from another package, pass the FQN string instead, for example `"builtin.core.knowledge.tools.query"`. TypeScript does not check these strings. Install resolves them and rolls the package back when one does not exist. The [Built-in catalog](../platform/built-ins.md) lists what CI ships, and the template pins the FQNs that it uses in a `bun test` that catches a typo before deploy.
44
44
 
45
45
  A package that declares Agents or Skills needs a person to install it; an install with no signed-in user fails. A package can depend on another package by FQN only, for example `dependencies: [{ fqn: "com.acme.base" }]`, and each install picks the latest eligible release of the dependency by promotion order.
46
46
 
@@ -61,6 +61,7 @@ GITHUB_TOKEN=$(gh auth token) opensrc path collective-intelligence-solutions/sdk
61
61
  - [Platform architecture](platform/architecture.md): The parts of Collective Intelligence, the three integration models, and how a package moves from build to install.
62
62
  - [Entrypoints](platform/entrypoints.md): The ways people, apps and agents enter CI at runtime, and the package subpaths that your code imports.
63
63
  - [Grants and permissions](platform/access.md): How CI decides what a user, a Group, an app or an API key may do, and how to change it in the template.
64
+ - [Built-in catalog](platform/built-ins.md): Every tool, workflow node, trigger, connector, schema, channel and MCP operation that CI ships, and which ones are already installed in an Organization.
64
65
  - Data resources
65
66
  - [Data resources](data/index.md): Choose between Knowledge, Datasets and Documents, and see where each kind keeps its data.
66
67
  - [Knowledge](data/knowledge.md): Declare schemas, store typed records, query them, and read and watch them from an external app.
@@ -0,0 +1,225 @@
1
+ # Built-in catalog
2
+
3
+ Every tool, workflow node, trigger, connector, schema, channel and MCP operation that CI ships, and which ones are already installed in an Organization.
4
+
5
+ Before you declare a tool, a workflow node, a connector or a schema in a demo package, check this page. CI may already ship it. Everything on this page is a package that ships with each CI version and uses an FQN under `builtin.` (schema templates use `core.`). Its manifest and declarations follow the same contracts as your own package, as [Extensions](../develop/extensions.md) describes. Your package refers to a built-in item by its FQN and lists its package as a dependency. Do not copy a built-in item into your package.
6
+
7
+ ## Install policies
8
+
9
+ A package that ships with CI is not always installed in an Organization. Each built-in package has one of three install policies:
10
+
11
+ | Policy | What happens | Packages |
12
+ | --------- | ---------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
13
+ | Mandatory | Installed in every Organization. An admin cannot uninstall it. | Core Workflows, Core Knowledge, Core Information, Core Activities |
14
+ | Standard | Installed when an admin creates an Organization with "Install packages" on, which is the default. An admin can uninstall it later. | Core Datasets, Core Documents, Core External Apps, Core Messaging, Core Notifications, Core OCR, Core Speech, Core Studio, Core Utilities, SQL databases |
15
+ | Optional | An admin installs it in each Organization that needs it. | Every integration in [Integrations](#integrations), Core Authoring, Core Reporting |
16
+
17
+ When CI is upgraded, it installs the new version of each mandatory package and of each optional package that an Organization already has. An upgrade never installs an optional package that the Organization did not have.
18
+
19
+ > **Warning:** When your manifest lists an integration such as `builtin.integrations.hubspot`
20
+ > in `dependencies` and the Organization does not have it, the deploy of your
21
+ > package installs it too. The integration still has no connection, so its tools
22
+ > and connectors fail until an admin creates one. When the Organization already
23
+ > has another release of the integration or of SQL databases, the deploy stops
24
+ > with a manual prerequisite. An admin then updates that package first.
25
+
26
+ When a package is installed, each of its tools gets the defaults that its declaration sets. With `autoInstantiate`, CI creates the tool in the Organization. With `autoActivate`, CI also turns it on. A tool that is created but not turned on appears in the Dashboard, and an admin turns it on. A tool without `usageScopes` can be used in chat, in Skills and in automation.
27
+
28
+ ## Workflow triggers
29
+
30
+ Core Workflows ships these triggers, with FQNs under `builtin.core.workflows.nodes.triggers.`:
31
+
32
+ | Trigger | Starts a run when |
33
+ | -------------------- | -------------------------------------------------------------------- |
34
+ | On-demand | A person or an app starts it |
35
+ | Schedule | A cron rule fires in the time zone that you set |
36
+ | Webhook | An authenticated `POST` arrives at its path; it can answer with JSON |
37
+ | Entity Created | A Knowledge record of the watched schema is created |
38
+ | Entity Updated | A Knowledge record changes; you can watch only some top-level fields |
39
+ | Entity Deleted | A Knowledge record is deleted |
40
+ | External application | An external application starts it through the server SDK |
41
+ | MCP Operation | An MCP client calls `invoke_operation` on it |
42
+
43
+ Integrations add message and file triggers. Slack, Telegram and WhatsApp add "message received". Twilio adds SMS received and WhatsApp message received. Microsoft 365 adds email received, Teams message received and Drive file created. Notion adds page created, page updated and page deleted. [Entrypoints](entrypoints.md) explains as which identity each trigger runs.
44
+
45
+ ## Workflow nodes
46
+
47
+ Core Workflows ships these action nodes, with FQNs under `builtin.core.workflows.nodes.actions.`:
48
+
49
+ | Collection | Nodes |
50
+ | ------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
51
+ | Control flow | If, Switch, Merge, Loop (each item or in batches; 100 iterations by default), Wait (a duration, a time, a callback or an event), Invoke Workflow, Stop & Error |
52
+ | AI | Prompt (a model step that returns text or JSON), Skill (runs one Skill) |
53
+ | Data | Set Variable, Adapter (filter, sort and reshape), Code Block, HTTP Request, Persist Artifact, Stored File Link, File Content, File URL |
54
+ | Human review | Approval |
55
+ | Workflow boundaries | Input, Output |
56
+
57
+ Some limits apply to these nodes. HTTP Request times out after 1 to 60 seconds (10 by default) and can read header values from Organization secrets. File Content reads up to 5 MiB. A File URL link lasts up to 7 days. Wait, Approval and Invoke Workflow pause the run durably and continue it later.
58
+
59
+ The standard packages add nodes for their own areas:
60
+
61
+ | Package | Nodes |
62
+ | ------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
63
+ | Core Knowledge | Create Knowledge Entity, Update Knowledge Entity, List Knowledge Entities, Save File to Knowledge (an HTTPS file up to 10 MiB), Add Account Knowledge Entity |
64
+ | Core Documents | Collection nodes (create, list, get, update, delete, restore), search nodes (Search Documents, Search Within Document, Expand Passage, Read Document Range), and sync nodes for connected collections |
65
+ | Core Datasets | Describe Dataset, Query Dataset Records, Query Dataset Aggregate, Query Dataset Time Series |
66
+ | Core Messaging | Send Email |
67
+ | Core Notifications | Send Notification |
68
+ | Core OCR | Extract Document |
69
+ | Core Speech | Transcribe Audio, Synthesize Speech |
70
+ | Core Utilities | Web Search, Generate Image |
71
+ | Core External Apps | Gateway Operation |
72
+
73
+ Each tool of an integration is also a workflow action node with the same name, so a workflow can call HubSpot or Gmail without a Skill.
74
+
75
+ ### Code Block runtimes
76
+
77
+ A Code Block runs Python or JavaScript in one of two runtimes:
78
+
79
+ | Runtime | Language |
80
+ | ----------------------- | ---------- |
81
+ | Python 3.12 | Python |
82
+ | JavaScript (Node.js 24) | JavaScript |
83
+
84
+ In isolated mode, each run starts a new container with no network, a read-only file system, 128 MB of memory and half a CPU. A run can take 5 seconds and write 16 KiB of output. Use a Code Block for small, deterministic data changes. Use HTTP Request for network calls.
85
+
86
+ ### Field types
87
+
88
+ Node settings use field types from two collections. Your own nodes can use them too.
89
+
90
+ - Standard fields: Text, Text list, Number, Integer, Money, Boolean, Select, Multi-select, Email, URL, UUID, JSON, Date, Time, Date and time, Date range, Cron, HTTP method, Code snippet, Condition builder.
91
+ - Platform fields: Connection, Entity schema, Entity instance, Model, Skill, Workflow, User, Users, File, Application Selector, Code Block Docker runtime.
92
+
93
+ ## Tools
94
+
95
+ These tools ship in the mandatory and standard packages. They are grouped in tool collections. "Ready" means that the tool is created and turned on when its package is installed.
96
+
97
+ | Tool collection | Tools | Ready |
98
+ | ----------------------- | --------------------------------------------------------------------------------------- | ----------------------------------------------------- |
99
+ | Knowledge | Knowledge Catalog, Knowledge Query, Knowledge Document Read, Knowledge Mutate | Yes, except Knowledge Mutate, which an admin turns on |
100
+ | Information | Advanced Information Retrieval, Expand Passage, Read Source Range | Yes |
101
+ | Documents | Documents Query | Yes |
102
+ | Datasets | Query Dataset (records, aggregates and time series; no SQL) | Yes |
103
+ | Web | Web Search (Tavily by default, or Exa), Website Scraper | An admin turns them on |
104
+ | Code Execution | Script Runner (short Python or Node.js calculations in a locked-down scratchpad) | An admin turns it on |
105
+ | Images | Generate Image | An admin turns it on |
106
+ | Messaging | Send Email | An admin turns it on |
107
+ | Notifications | Send Notification | Yes |
108
+ | OCR | Extract Document (PDF, PNG, JPEG or WebP to Markdown, or to JSON that matches a schema) | An admin turns it on |
109
+ | Speech | Transcribe Audio, Synthesize Speech | An admin turns them on |
110
+ | External App Run Events | Publish Run Event | Yes |
111
+
112
+ Two optional packages add more tools after an admin installs them. Core Reporting adds Build Report, Run Report, Create Report Definition, Update Report Definition, Archive Report Definition, List Report Definitions, List Report Runs and Read Report Run. Core Authoring adds the tools of the Authoring Assistant.
113
+
114
+ [AI](../ai/index.md) explains how to bind a tool to a Skill or an agent, and [Knowledge](../data/knowledge.md) explains which Knowledge tool to choose.
115
+
116
+ ### Authoring Assistant
117
+
118
+ Core Authoring ships one agent, the Authoring Assistant. From a conversation, it lists the catalog of tools, Skills, nodes and connections that the person may use, creates private drafts of agents, Skills and workflows, edits them, runs pre-publish review and Sandbox Runs, reads run traces, imports a `SKILL.md` folder and turns a plain-language schedule into a Schedule trigger. It never publishes, turns on, shares or approves anything. A person does those steps.
119
+
120
+ ## Data
121
+
122
+ [Data resources](../data/index.md) explains Knowledge, Datasets and Documents. This section lists the parts of them that CI ships.
123
+
124
+ ### Knowledge connectors
125
+
126
+ | Connector | Package | What it syncs |
127
+ | ----------------- | ------------------ | -------------------------------------------------------------------------------------------------------------------- |
128
+ | Structured File | Core Knowledge | CSV, XLSX and JSON rows into your schemas, from an upload, an FTP server, Google Drive or Microsoft 365. Import only |
129
+ | Knowledge gateway | Core External Apps | Records in both directions with an external application's server |
130
+ | PostgreSQL | SQL databases | Selected tables into Knowledge. Import only |
131
+ | MySQL | SQL databases | Selected tables into Knowledge. Import only |
132
+ | HubSpot API | HubSpot | Contacts, companies and deals into the CRM schemas. A read-only mirror |
133
+ | monday.com API | monday.com | Workspaces, boards, groups, items and updates into the project management schemas. A read-only mirror |
134
+ | Notion API | Notion | The Notion databases that you choose, into any schema. Import only |
135
+
136
+ No connector imports on its own schedule. After an admin creates a connector instance, they start an import or set an import schedule. During setup, HubSpot and monday.com suggest a delta import every 15 minutes and a full import each night, and Notion suggests a delta import every 5 minutes. [Connectors](../data/connectors.md) explains bindings, sync modes and mapping.
137
+
138
+ ### Knowledge schemas
139
+
140
+ CI ships schema templates in domains, with FQNs under `core.schemas.<domain>.`. Your custom schemas can point a relation field at them, so a demo does not need its own Company or Person schema.
141
+
142
+ | Domain | Schemas | When they appear |
143
+ | ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------- |
144
+ | Strategy | Organization context, strategic dimension, business goal, strategic initiative, metric, constraint, assumption, decision principle, stakeholder, context note | When an Organization is onboarded |
145
+ | Directory | Organization, organizational unit, group, person | When an Organization is onboarded |
146
+ | Assets | Asset | When an Organization is onboarded |
147
+ | CRM | Contact, company, deal, lead, pipeline, stage, activity, task, note | When you choose "Create from template" while mapping a connector |
148
+ | HR | Worker, department, job opening, job opening template, candidate, application, interview, contract, time off, performance review | When you choose "Create from template" while mapping a connector |
149
+ | Project management | Workspace, project, group, task, update | When you choose "Create from template" while mapping a connector |
150
+
151
+ A template creates a schema only once. A later CI version never changes a schema that already exists, so the edits that an admin makes stay.
152
+
153
+ ### Documents and Datasets
154
+
155
+ [Documents](../data/documents.md) lists the two Documents connectors, Google Drive Documents and Microsoft 365 Documents. [Datasets](../data/datasets.md) lists the Dataset providers: managed PostgreSQL storage in CI, and external Databricks, Microsoft Fabric and Remote Gateway sources. A managed import reads CSV, JSON, JSON Lines, Parquet and Excel (the first worksheet).
156
+
157
+ ### Connection types
158
+
159
+ A connection holds the address and credential of one external system. CI ships these connection types:
160
+
161
+ - Core Knowledge: FTP server.
162
+ - SQL databases: PostgreSQL database, MySQL database.
163
+ - Core External Apps: External app gateway.
164
+ - Core Messaging: Resend, SendGrid, Amazon SES and Azure Communication Services, for email.
165
+ - Each integration: the connection that its row in [Integrations](#integrations) names.
166
+
167
+ ## Integrations
168
+
169
+ All integrations are optional. The admin needs the credential in the last column before they install one.
170
+
171
+ | Integration | What it adds | Credential |
172
+ | ---------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------- |
173
+ | Apollo | 64 tools in three families: prospecting and enrichment, accounts, contacts, tasks and lists, and sequences and outreach email | API key |
174
+ | Lusha | 13 tools that search, enrich and prospect contacts and companies | API key |
175
+ | HubSpot | 17 CRM tools for objects, lists, associations, owners, pipelines and properties; the HubSpot API Knowledge connector | Private app access token |
176
+ | monday.com | The monday.com API Knowledge connector. No tools | API token |
177
+ | Notion | 8 tools that read, search, create, update and delete pages and databases; page triggers; the Notion API Knowledge connector | Internal integration token |
178
+ | Google Workspace | 23 tools for Gmail, Calendar and Drive; the Google Drive Documents connector; sending email from a mailbox; Google sign-in | Service account with domain-wide delegation |
179
+ | Microsoft 365 | 24 tools for Outlook mail, Calendar, OneDrive and Teams; mail, Teams and Drive triggers; the Microsoft 365 Documents connector; sending email from a mailbox; Microsoft Entra sign-in | Entra application (client secret, certificate or federated), and a Teams bot for Teams messages |
180
+ | Slack | Send Slack message, and a message trigger | Bot token and signing secret |
181
+ | Telegram | Send Telegram message, and a message trigger | Bot token |
182
+ | WhatsApp | Send WhatsApp message, and a message trigger | Meta Cloud API access token, app secret and business account ID |
183
+ | Twilio | Send SMS and Send WhatsApp message, with matching triggers | Account SID and Auth Token |
184
+ | Databricks | A read-only Dataset provider for Unity Catalog through one SQL Warehouse | Service principal; personal access tokens are refused |
185
+ | Microsoft Fabric | A read-only Dataset provider for Warehouses and Lakehouse SQL endpoints | Entra service principal; personal access tokens are refused |
186
+
187
+ In Apollo, Lusha and HubSpot, tools that only read can be used in chat, in Skills and in automation, and tools that write can be used only in Skills and in automation. Apollo enrichment and Lusha search tools count as writes because each call spends credits on the provider's account. The message tools of Slack, Telegram, WhatsApp, Twilio and Teams can be used only in Skills and in automation. The other Google Workspace, Microsoft 365 and Notion tools can be used everywhere, so an Organization that offers them in chat lets a model send mail or edit pages from a conversation. WhatsApp sends only a template message to a person who has not written in the last 24 hours.
188
+
189
+ ## Email and notifications
190
+
191
+ Send Email and Send Notification deliver through what an admin configures:
192
+
193
+ - Email providers: Resend, SendGrid, Amazon SES and Azure Communication Services. Google Workspace and Microsoft 365 add sending from a mailbox.
194
+ - Notification channels: Studio Internal, which CI creates in every Organization, Email, Slack webhook, Microsoft Teams webhook, Discord webhook and Telegram bot.
195
+
196
+ ## MCP
197
+
198
+ CI is an MCP server named `collective-intelligence-mcp`, at `/mcp` and at `/mcp/o/<organizationId>`. It offers these tools:
199
+
200
+ | Tool | What it does |
201
+ | ------------------ | ------------------------------------------------------ |
202
+ | `list_memberships` | Lists the Organizations that the caller belongs to |
203
+ | `list_skills` | Lists the Skills that the caller can run |
204
+ | `run_skill` | Starts a Skill run in the background |
205
+ | `get_skill_run` | Reads the status and result of a Skill run |
206
+ | `list_operations` | Lists the workflows that have an MCP Operation trigger |
207
+ | `invoke_operation` | Starts one of those workflows and returns a handle |
208
+ | `get_operation` | Reads the status and result of an operation |
209
+ | `cancel_operation` | Cancels a running operation |
210
+ | `explain_access` | Explains why a principal can or cannot do something |
211
+
212
+ The Dashboard has ready client settings for Claude (OAuth) and Codex (bearer token). Two MCP Profiles ship. Admin full-access is for administrators and has no rate limits. User-context is for other users and allows 5 operations at a time, 60 starts a minute and 10,000 starts a month. CI does not connect to other MCP servers, and no integration adds one. [Entrypoints](entrypoints.md) explains MCP sign-in and tokens.
213
+
214
+ ## Studio and models
215
+
216
+ Core Studio ships the Default Studio app. Its AI section has Activity Center, Chat, Approvals, Notifications, Agents and Skills, and its Productivity section has Documents. [Studio](../ai/studio.md) lists the built-in views that your own Studio app can use.
217
+
218
+ CI ships a catalog of models from several providers. Each Organization chooses a model preset, and the preset sets one model for each tier (complex, balanced, fast and super fast) and for embeddings, OCR, speech-to-text and text-to-speech. A package asks for a tier, not a model. The Prompt node uses the balanced tier. [AI](../ai/index.md) explains how a Skill chooses its model.
219
+
220
+ ## Next steps
221
+
222
+ - [Extensions](../develop/extensions.md)
223
+ - [AI primitives](../ai/index.md)
224
+ - [Connectors](../data/connectors.md)
225
+ - [Choosing primitives](../scenarios/index.md)
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@collective-intelligence/external-apps-server-sdk",
3
- "version": "0.3.8",
3
+ "version": "0.3.9",
4
4
  "description": "Server-side client and proxy adapters for Collective Intelligence external applications.",
5
5
  "type": "module",
6
6
  "sideEffects": false,
@@ -151,8 +151,8 @@
151
151
  "test:postgres": "bun tools/test-postgres.ts src/adapters/persistence/drizzle"
152
152
  },
153
153
  "dependencies": {
154
- "@collective-intelligence/extensions-sdk": "0.3.8",
155
- "@collective-intelligence/external-apps-client-sdk": "0.3.8",
154
+ "@collective-intelligence/extensions-sdk": "0.3.9",
155
+ "@collective-intelligence/external-apps-client-sdk": "0.3.9",
156
156
  "ajv": "8.20.0",
157
157
  "ajv-formats": "3.0.1",
158
158
  "jose": "6.2.9"
@@ -251,5 +251,5 @@
251
251
  "vitest": "5.0.2",
252
252
  "zod": "4.4.3"
253
253
  },
254
- "gitHead": "d7c921799df228a99d1adbb7ba8be2010460e445"
254
+ "gitHead": "56627fcab8f5db0afbfc57347d0887e6cf7f12a9"
255
255
  }