Access Key field configuration

Access Key passthrough configuration

Access Key usage mode

Enumeration values:

• 1: Access Key pass-through

Access Key field type. 1: AccessKeyId, 2: AccessKeySecret, 3: SessionToken

Required or not

header/query field name

Default value of the AccessKey, which can be empty.

Unified injection position of the Access Key field. 0: Header, 1: Query

Employee sub-account id

Employee nickname

Employee avatar

Maximum reasoning round

Agent ID

Agent Basic Configuration

System prompt

Model information

Tool details

Plug-in configuration

Skill details.

Advanced configurations.

List of Function Tools executed by the caller

Input parameter limit: only available in client user state Agent scenario. This field is ignored in business configuration state Agent

Tool type

Input parameter limit: currently only support "function"

Tool name

Tool description

Tool input parameter definition

Input source type: 0 user input, 3 custom variable (API parameter)

User handwriting input

System parameter

Custom Variable (API parameter)

Environment variable parameter

Application variable parameters

Unique model id

Model alias

Model context length character limit

Instruction length character limit

Model parameters

Basic plug-in configuration

Plugin name.

Plug-in icon url

Plugin description

Plug-in product type

Enumeration values:

• 0: ordinary plug-in
• 1: connector plug-in

Plug-in status

Enumeration values:

• 0: unknown
• 1: available
• 2: unavailable

Plug-in authentication configuration status

Enumeration values:

• 0: No authorization required
• 1: Unconfigured
• 2: Configured

plugin_id

Plug-in Header parameter

Plug-in Query parameter

Whether to use CAM one-click authorization. This parameter is valid only when auth_type=2

Authorization type

Enumeration values:

• 0: No authentication
• 1: API Key
• 2: CAM authorization
• 3: OAuth2.0 authorization

OAuth authorization and consent mode; 0 - developer authorization; 1 - user authorization (valid only when auth_type=3)

Credential configuration

Plugin authentication value source

Enumeration values:

• 0: Unspecified
• 1: Use the default authentication value of the plug-in. Only APIKey/AccessKey is supported.
• 2: Refer to a credential
• 3: Refer to a variable

Credential ID

Input parameter limit: required when AuthValueSource=2

Parameter configuration

Parameter location

Enumeration values:

• 0: Header authentication
• 1: Query authentication

Parameter name

Source of parameter value

Parameter name

Required or not

Input value

Agent Name

Icon URL

Agent role: 0=primary / 1=secondary

Agent description

Application name.

Developer

Primary AgentId, read-only, cannot be updated through API modification

Name.

Update time, unix timestamp in seconds (s)

Action description.

Change to test

AgentID

Agent Name

Update time, unix timestamp in seconds (s)

Status, status value: 1: Add new, 2: Modify, 3: Delete

Action description.

Publish a message

Release detail

skillId

skill name

Skill description

skill display name

Skill show description

skill icon url

Skill source

Skill version

Skill ID

Agent Basic Configuration

System prompt

Primary model configuration

Tool information

Plugin information.

Skill information.

Advanced settings

List of Function Tools executed by the caller

Input parameter limit: only available in client user state Agent scenarios. B-end configuration state Agent ignores this field and

AgentId

Agent Identity Portrait

Advanced setting; returned when scope=0

System parameter name

Configuration of dialogue history rounds; used only when Input is "dialogue history" in system variables;

Tool configuration field

Tool name

Tool status

Enumeration values:

• 1: available
• 2: unavailable
• 3: invalid

Calling method

Enumeration values:

• 0: Non-streaming
• 1: Streaming

Tool access mode

Enumeration values:

• 0: Not specified
• 1: Read-only
• 2: Write/Delete

plugin_id

Tool id

Description.

Tool input parameter list

Tool output parameter list

Tool Header parameter list

Tool Query parameter list

Tool source: 0-comes from plug-in, 1-comes from workflow

Whether to disable

Tool configuration

Tool name

Tool description

Tool parameter type

Enumeration values:

• 0: STRING
• 1: INT
• 2: FLOAT
• 3: BOOL
• 4: OBJECT
• 5: ARRAY_STRING
• 6: ARRAY_INT
• 7: ARRAY_FLOAT
• 8: ARRAY_BOOL
• 9: ARRAY_OBJECT
• 20: ARRAY_ARRAY
• 99: NULL

Required or not

Sub-parameter, used only when the data type is OBJECT or ARRAY<>

Whether to hide from the model in mode

Parameters in types of OneOf

Parameters in types of AnyOf

Source of parameter value

Parameter name

Variable description

Parameter type

Enumeration values:

• 0: STRING
• 1: INT
• 2: FLOAT
• 3: BOOL
• 4: OBJECT
• 5: ARRAY_STRING
• 6: ARRAY_INT
• 7: ARRAY_FLOAT
• 8: ARRAY_BOOL
• 9: ARRAY_OBJECT
• 20: ARRAY_ARRAY
• 99: NULL

Sub-parameter, used only when the data type is OBJECT or ARRAY_OBJECT

Parsing method

User input parameter value

Request body parameters

Example

External call address of API plug-in

Header

Request method

Output

Query parameter

Address

Auxiliary information (sub-status/approval/appeal/search for resource/special status)

Configuration.

Metadata

Application key information

Sharing link information (including access control)

Status.

Business shared configuration

List of shared knowledge bases referenced by the application

portal shared configuration

Application-mounted cloud disk

Whether to enable context rewrite

Whether to enable image and text retrieval.

Reply flexibility

Custom configuration on the dialogue side (shared across all modes, with permission to modify configuration dynamically during a dialogue)

Intent attainment priority

Application cloud disk configuration

1: Add new 2: Modify 3: Delete

Welcome message content

Opening problem list

Application ID.

Application name.

Application mode

Enumeration values:

• 1: Standard mode
• 2: agent mode
• 3: Single workflow mode
• 4: ClawAgent mode

Whether to enable access control

Enumeration values:

• true: Enable
• false: Disable

Whitelist information

Allowlist array information

Parameter format: allowlist value

Application status

Enumeration values:

• 1: not launched
• 2: running
• 3: disabled
• 4: importing

Status description.

Application ID.

Application mode. Enumeration values: 1: standard mode, 2: Agent mode, 3: single workflow mode, 4: ClawAgent mode

Application avatar

Application name.

Operation information

Status information

Sub-status information

Resource operation permissions

Application description information

Input parameter

Output parameters

Application ID.

Execution configuration

Number of failures

Push channel configuration

Number of successes

Trigger configuration

Trigger ID

Trigger name

Trigger status

Visitor ID

Application id

Session id

Trigger creation time

End time.

Trigger running instance id

Request ID

Result code

Result Summary

Single dialogue id

Trigger start execution time

TraceId, used for logging

Trigger id

Visitor ID

Workflow running ID

Parameter name.

Parameter type

Enumeration values:

• 0: string
• 1: integer
• 2: floating-point number
• 4: object
• 5: string array
• 6: integer array
• 7: floating-point number array
• 8: boolean value array
• 3: boolean value
• 9: object array
• 10: file
• 11: document
• 12: image
• 13: audio
• 14: video
• 15: file array
• 16: document array
• 17: image array
• 18: audio array
• 19: video array
• 20: nested array
• 22: key
• 99: null
• 100: unspecified type, for OneOf and AnyOf scenarios

Parameter value

Bind parameter list

Parameter value

Application variable name

Parameter name

Required or optional

Sub-parameter list

Trigger execution prompt content

api parameter binding

Session id

Execution time

End time

Parameter format: YYYY-MM-DD hh:mm:ss

Trigger instance id

Result code

Result Overview

Single dialogue id

Trigger time

Parameter format: YYYY:MM:DD hh:mm:ss

Execution start time

Parameter format: YYYY:MM:DD hh:mm:ss

Trigger id

Whether it has been read

Visitor Id

Workflow run id

Trigger scheduled configuration.

Last trigger time

Parameter format: YYYY-MM-DD hh:mm:ss

Next trigger time

Parameter format: YYYY-MM-DD hh:mm:ss

Trigger mode

Application ID.

Number of failures

Last session id

Value range: [0, 2]

Number of successes

Trigger id.

Trigger name

Trigger execution status

The number of unread logs

Visitor ID.

Trigger webhook parameter configuration

key of the webhook

webhook key

webhook domain name or IP address

Trigger API parameter list

Push Webbook address

Workflow API parameter binding

Workflow parameter binding

Workflow ID

Workflow name

Employee information

Application business id

Application name

Name when the operation log is triggered

Operation time

Parameter format: timestamp in seconds

Operation type.

Operation object.

Operation content

Unique ID for the operation.

Authorization mode.

Enumeration values:

• 0: No authentication
• 1: API Key authentication
• 2: CAM authorization
• 3: OAuth 2.0 authorization

API Key Authorization Configuration

CAM Authorization Configuration

OAuth 2.0 authorization configuration

AccessKey authorization configuration

Cash price

Unit: CNY

PU price

Unit: pu

Property name

Attribute value.

Caller ID, whose meaning is determined by subject_type (such as app_id, kb_id)

Name of the calling entity

Caller type: APP/KB/WIDGET/OPEN_CLAW/KB_RECALL_TEST/WORKBENCH/MODEL_API

Role name.

Key location HEADER/QUERY

Enumeration values:
| uint | Description |
| --- | --- |
| 0 | Header authentication |
| 1 | Request information authentication |

SecretId field name

SecretKey field name

CAM Access Key Field Configuration

Whether CAM role authorization is supported.

Category ID

List of category IDs in the path from the root node

List of category names in the path from the root node

Whether the current user can add subcategories

Whether the current user can delete the category

Whether the current user can edit this category

Whether to enable Agent team collaboration

prompt content

Whether to enable channel context timeout clearing

Channel context timeout duration

Value range: [1, 99]

Agent Team Collaboration Configuration

Long-term memory configuration

Channel context timeout configuration clearing

Whether to enable long-term memory

Code

Example.

Input parameter

Output parameters

Composite Billing List

Composite billing dimensions

Cash price

Unit: CNY

pu price

Unit: pu

Call source (subject_type determines the meaning of subject_id/subject_name. For example, for APP, subject_id=app_id and subject_name=app_name)

Time when the limit was exceeded (Unix seconds)

Model name

Request content (the original query text of the user request)

Space ID

Answer

Knowledge application scope: 1=disabled, 2=development domain only, 3=release domain only, 4=global

Associated document name.

Associated document type

QA ID

Query

Source type: 1=documentation generation, 2=batch import, 3=manual entry

Update time (Unix seconds, for sorting old and new)

Conflict group ID

App ID

Session ID

Creation time.

Session type enumeration values: 0-CONVERSATION_TYPE_UNSPECIFIED (unspecified; means all in list query), 1-CONVERSATION_TYPE_VISITOR (visitor trial), 2-CONVERSATION_TYPE_EVALUATION (evaluation), 5-CONVERSATION_TYPE_API (API integration), 10-CONVERSATION_TYPE_WORKFLOW (workflow debugging), 20-CONVERSATION_TYPE_SHARE (share a link)

Update time.

Session title

Client-side AgentId used by the session

Task content

Task number

Task status. pending: to be executed, processing: processing, success: completed, failed: processing failure, stop: canceled

Text content

Content type. text: text, image: image, file: file, custom_variables: custom input parameter information, widget_action: widget action information

Custom parameter data

Custom parameter data

Custom variable data

Business form

Tab List

Tab list.

Tab mode. Enumeration values: 0-OPTION_MODE_SINGLE (single selection), 1-OPTION_MODE_MULTI (multiple selection)

Reference marker information list

List of reference superscript information

Reference source list

Reference list

Associated record ID

Intelligent Agent Task List

Intelligent Agent Task List

Workflow input parameter

MCP-APP call information

AI Call Configuration

Background image configuration

Fallback response switch

Whether to use recommended questions.

Whether to use online search.

Fallback response

Input box configuration

Output method. Enumeration values: 1: streaming, 2: non-streaming

Recommended question generation prompt mode. Enumeration value: 1: Prompt that outputs recommended questions combined with the knowledge base only

A new message interrupts the ongoing message

Maximum number of files

Supported file type list

Supported file type list

Capacity boundary: A single request is read-only for resources belonging to this plugin

ui:// resource, based on which the frontend calls ReadMCPResource to pull HTML

thread on the agent-exec side

JSON: The complete original CallToolResult for replay when rebuilding historical sessions

Session ID.

Message icon

Message ID

Message name

Record ID

Message role

Message status. pending: pending execution, processing: in processing, success: completed, failed: processing failure, stop: canceled

Status description

Message title

List of message contents

Message content list

Type.

Index value of the reference source

Reference source location

Failure error code of the dialogue

Dialogue failure error message

Reply record ID, corresponding to the record_id of the reply message in messages

ID of the user's question record, corresponding to the record_id of the user message in messages

Duration info of a single dialogue

token consumption information of a single dialogue

Description of a single dialogue failure; empty if successful

Status of a single employee assistant dialogue.

Enumeration values:

• pending: Pending,
• processing: Processing,
• success: Success,
• failed: Failure,
• stop: Stopped.

Total time consumed for a single dialogue, in milliseconds

Time to first token, in milliseconds

Total time consumed for model inference, in milliseconds

Total input token count

Total output token count

Total token consumption

Total number of cache hit tokens

Total number of reasoning tokens

Reference index

Reference name

REFERENCE source TYPE. Enumeration values: 0-APP_REFERENCE_TYPE_UNSPECIFIED (not specified), 1-APP_REFERENCE_TYPE_QA (QA), 2-APP_REFERENCE_TYPE_SEGMENT (fragment), 3-APP_REFERENCE_TYPE_DOC (document), 4-APP_REFERENCE_TYPE_WEB_SEARCH (WEB SEARCH), 5-APP_REFERENCE_TYPE_GRAPH_RAG (GraphRAG)

Last reset millisecond-level timestamp

Last boundary reset; this record and earlier records are no longer used as dialogue context

Base64-encoded original widget data

payload from the user's last submission

Location of the widget in the text

widget status JSON data

widget view data

component_ID

Component operation ID

Workspace ID

Storage type

Business shared switch

Share scope type. 1: all business staff, 2: designated account, 3: designated space

Business shared application tag

Shared range information (for users, StrId is uin and Name is username; for spaces, StrId is space ID and Name is space Name)

Whether to enable

Whether to enable dynamic configuration modification on the dialogue side

Count of calls for this reply method

Percentage of this reply method in the total number of calls (percentage, 0–100, rounded to two decimal places; all 0 when no data is available, and the sum of all items is 100 when there is data)

Reply method name (internationalized according to the request language; falls back to the Enumeration Name of reply_method when i18n is missing)

External URL

Whether to replace the original display

Confidence threshold

Whether to enable

Return the top N entries

Whether it is downloadable

Whether to show in the reference

Input cash price

Unit: CNY

Input PU price

Unit: pu

Output cash price

Unit: CNY

Output PU price

Unit: pu

Whether storage method modification is supported

ES password (encrypted)

ES cluster ID

Storage type: 1=default storage, 2=custom storage

ES userName

Validity end time (Unix seconds. 0 indicates permanent validity).

Validity start time (Unix seconds)

Prompt configuration

Workflow configuration.

Whether to enable timeliness retrieval. Once enabled, retrieval results will be sorted based on the validity time of the knowledge

Validity time

Behavior after the validity period expires: 1=NOT_RETRIEVABLE cannot be retrieved, 2=RETRIEVABLE can still be retrieved; meaningless for permanent validity

Parameter name

Parameter format: path of the specified field to get

Access mode: RO=read-only, RW=read-write; required. Only RW can be used when mounting to /workdir.

Absolute mount directory in the sandbox; /workdir can be mounted at most once and only RW can be used.

Cloud disk identification

Server-side generated mount flag. Reserved for modification, optional for new additions.

Path relative to the root directory of cloud disk data; an empty string indicates mounting the entire data root directory.

Cloud disk name. It is supplemented by the server only when querying the application configuration and ignored during modification. It is empty when it cannot be obtained and does not impact the mount configuration.

Model name

Whether to enable

Whether it can be added

Whether it is deletable

Whether it is editable

Category ID

Subcategory list.

Is leaf node (no subcategory)

The number of category objects

Metadata configuration (configuration message when this category is set as metadata)

Category name

Embedding model

QA extraction model

Schema generative model

Scaling operation: 1=trigger recovery from exceeded state

Database search configuration

Document retrieval configuration.

Knowledge base ID

Retrieval optional configuration

QA search configuration

ReRank configuration

Search filter configuration

Retrieval strategy: 1=hybrid, 2=semantic, 3=keyword, 4=none

Text2SQL model

Description.

ES configuration

Model configuration

Knowledge base name

Owner ID

Confidence degree

Hit knowledge type: 1=QA, 2=document fragment, 3=database, 4=map

Recall type list

Result payload

Result type

Additional information for similar questions.

Basic information of knowledge fragment

Knowledge source information

Effective scope

Applicable scope (tag condition).

Content

Associated document ID

Document name

Knowledge base ID

Knowledge ID

Query

Document title

Whether it is big data

Whether it is a shared knowledge base

Knowledge base name

Tag name.

List of incremental modifications to standard words (add/modify/delete)

Tag ID.

Tag name.

Tag standard term ID list.

Tag standard word list

Tag ID

Tag standard term ID list

Tag references list.

Tag ID.

Metadata configuration (configuration message when the tag is set to metadata)

Tag name.

Number of resources referring to the tag

Tag value (standard word + synonym list)

Total number of tag values.

Synonym list

Standard words

Standard term ID (generated by the backend, not passed in upon creation)

Operation type (not allowed to be 0. Value: 1=create, 2=modify, 3=delete)

Synonym list (pass the complete synonym collection during CREATE and UPDATE, overriding update)

Standard word (required for CREATE and UPDATE, can be left blank for DELETE)

Standard term ID (required for UPDATE and DELETE; leave blank for CREATE, generated by the backend)

MCP plug-in public access address

MCP server address

MCP transmission type: SSE/Streamable
Enumeration values:
| uint | Description |
| --- | --- |
| 0 | SSE + HTTP mode |
| 1 | Streamable HTTP mode |

header parameters of the MCP plug-in

query parameter of the MCP plug-in

SSE persistent connection timeout period, in seconds

Request timeout period, in seconds

Authorization information.

Whether the interactive interface (MCP Apps) and plug-in-level tags are supported. Default false

Input parameter

Output parameters

Tool meta information

Whether the interactive interface (MCP Apps) and plug-in-level tags are supported. Default value: false

UI extended meta information of the tool, corresponding to _meta.UI in the MCP protocol, declaring the interactive UI resource (ResourceUri) associated with the tool and the caller Visibility. Returned only when the tool supports MCP Apps or declares Visibility; this field is empty for text tools. For details, see the MCPToolUIMeta structure definition.

Associated UI resource URI in UI:// scheme, format: UI://<plug-in flag>/<resource name>-<version>. This field is the entry to the interactive UI of MCP Apps. If not empty, it means the tool supports Apps (shown as "text + interactive UI"). If empty, it is a plain text tool. It is automatically identified and filled based on the tool sync result and does not support manual editing.

Visibility declaration for tool callers. Value range: model (callable by models), app (callable by UI). Multiple selections allowed, for example ["model","app"]. Independent of ResourceUri (SEP-1865) and can exist independently. For example, a backend app-only tool is ["app"]. When ResourceUri is non-empty and this field is not specified, it is normalized to ["model","app"] by default. Existing non-Apps tools remain empty.

Enumeration values:

• model: Supports model
• app: Supports app

Metadata value name (for display only).

Whether to refer to all values of this type (when true, ref_value_id should be 0)

Business ID referenced by metadata (property ID, category ID, etc.); this field should be 0 when ref_all=true

Metadata usage scenario: 1=retrieve only, 2=both retrieve and generate

Metadata value type: 1=property tag, 2=document classification, 3=QA classification

Model badge list

Model author information

Model restrictive information

Model basic information

Model hyperparameter configuration

Model attribute configuration

Model provider info

Model status information

Model tag list

Display copy (rendered as final copy)

Style theme

tooltip copywriting (rendered as final copywriting). If it is empty, it is not displayed.

Badge type

Model description

Model icon address

Model ID

Model type

Model name

Author alias.

Author flag

Model context length display copy (for example, 128K, 1000K)

Character limit for model dialog box input length

Maximum input Tokens display copy (for example, 960k, formatted on the server, empty string if not configured)

Display copy for the maximum output Tokens (for example, 64k, formatted on the server, empty string if not configured)

Model prompt content length character limit

Default value.

Available Option List

Maximum value (valid only for numerical type).

Minimum value (valid only for numerical types).

Hyperparameter name

Hyperparameter type. 1-floating-point number, 2-integer, 3-string

Whether to enable deep thinking

Frequency penalty

Maximum output length

Existence penalty

Deep thinking effect

Repetition penalty

Output format (text, json_object)

Random seed

Stop sequence

Temperature

top_p

top_k

Property name

Attribute value.

Model provider alias

Model provider name

Model provider type

Number of exclusive concurrencies

Whether it is exclusive concurrency

Whether direct experience in the intelligent workbench is supported (model plaza details page: dynamically determined based on space_id; when false, the frontend disables the "experience now" button)

Resource status

RPM call cap (current account dimension: use the configured value if available; otherwise, take default. 0 means not obtained)

TPM call upper limit (at the current account level: use the configured value if available; otherwise, take default. 0 means failed to obtain)

Number of calls (top-level count of API calls)

Whether it is the default knowledge base

Model name, flag indicating the AI model used

MODEL domain consumption measurement list (authoritative field): lists each measurement category by unit + label. When unit=TOKEN, label is case-sensitive for Token subcategories (input/output/avg_/cache_), with an empty label indicating total_tokens; unit=PAGE_COUNT refers to the model's consumed page count

Total model PU consumption (sum of PU consumption within the aggregate dimension)

Business ID of the category

Subcategory list, tree nesting

Category name

Operation permissions of the current user on this category.

The number of message records in this category

Node running information

Node operation content.

Node operation process

Node operation logs

Statistical information.

Batch operation summary

End time (timestamp in milliseconds).

Start time (timestamp in milliseconds)

Logs of node.

URLs of the complete content of logs of node (has a value only when the Log content exceeds the limit)