@celigo/api-specs 0.2.19 → 0.2.20
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/dist/ai-agent.yml +308 -36
- package/dist/connection.yml +65 -10
- package/dist/flow.yml +18 -4
- package/dist/guardrail.yml +263 -20
- package/dist/import.yml +314 -40
- package/dist/job.yml +168 -110
- package/dist/sync.yml +323 -56
- package/dist/tool.yml +321 -45
- package/package.json +1 -1
- package/schemas.json +1997 -1882
package/dist/guardrail.yml
CHANGED
|
@@ -584,14 +584,22 @@ components:
|
|
|
584
584
|
AI Agent configuration used by both AiAgentImport and GuardrailImport (ai_agent type).
|
|
585
585
|
|
|
586
586
|
Configures which AI provider and model to use, along with instructions, parameter
|
|
587
|
-
tuning, output format, and available tools.
|
|
587
|
+
tuning, output format, and available tools. Providers come in two families with
|
|
588
|
+
two configuration formats:
|
|
588
589
|
|
|
589
|
-
- **
|
|
590
|
-
|
|
591
|
-
|
|
590
|
+
- **Built-in providers** — each has its own configuration block.
|
|
591
|
+
- **openai**: OpenAI models (GPT-4.1, GPT-5, etc.). Configure via the `openai` object.
|
|
592
|
+
- **gemini**: Google Gemini models via the LiteLLM proxy. Configure via `litellm` with overrides in `litellm._overrides.gemini`.
|
|
593
|
+
- **anthropic**: Anthropic Claude models via the LiteLLM proxy. Configure via `litellm` with overrides in `litellm._overrides.anthropic`.
|
|
594
|
+
- **Catalog providers** — `mistral`, `xai`, `huggingface`, `deepseek`, `cohere`, `groq`. Added
|
|
595
|
+
as catalog data rather than platform code, they share one generic flat format:
|
|
596
|
+
`model`, `modelOptions`, `instructions`, `output`, and the `tools` / `prompts` /
|
|
597
|
+
`resources` arrays directly on this object.
|
|
592
598
|
|
|
593
|
-
A `_connectionId`
|
|
594
|
-
|
|
599
|
+
A `_connectionId` on the parent import selects bring-your-own-key (BYOK)
|
|
600
|
+
credentials. Without one, built-in providers run on platform-managed
|
|
601
|
+
credentials; catalog providers have no platform-managed credentials, so they
|
|
602
|
+
save without a connection but cannot run until one is attached.
|
|
595
603
|
required:
|
|
596
604
|
- provider
|
|
597
605
|
properties:
|
|
@@ -601,18 +609,224 @@ components:
|
|
|
601
609
|
- openai
|
|
602
610
|
- gemini
|
|
603
611
|
- anthropic
|
|
612
|
+
- mistral
|
|
613
|
+
- xai
|
|
614
|
+
- huggingface
|
|
615
|
+
- deepseek
|
|
616
|
+
- cohere
|
|
617
|
+
- groq
|
|
604
618
|
x-enumDescriptions:
|
|
605
|
-
openai: Use OpenAI models via the OpenAI Responses API.
|
|
606
|
-
gemini: Use Google Gemini models via the LiteLLM proxy.
|
|
607
|
-
anthropic: Use Anthropic Claude models via the LiteLLM proxy.
|
|
619
|
+
openai: Use OpenAI models via the OpenAI Responses API (built-in; `openai` block).
|
|
620
|
+
gemini: Use Google Gemini models via the LiteLLM proxy (built-in; `litellm` block).
|
|
621
|
+
anthropic: Use Anthropic Claude models via the LiteLLM proxy (built-in; `litellm` block).
|
|
622
|
+
mistral: Use Mistral models (catalog provider; generic flat format, BYOK).
|
|
623
|
+
xai: Use xAI Grok models (catalog provider; generic flat format, BYOK).
|
|
624
|
+
huggingface: Use Hugging Face inference endpoint models (catalog provider; generic flat format, BYOK).
|
|
625
|
+
deepseek: Use DeepSeek models (catalog provider; generic flat format, BYOK).
|
|
626
|
+
cohere: Use Cohere models (catalog provider; generic flat format, BYOK).
|
|
627
|
+
groq: Use Groq-hosted models (catalog provider; generic flat format, BYOK).
|
|
608
628
|
description: AI provider to use.
|
|
609
629
|
x-celigo-ai-guidance:
|
|
610
630
|
- |-
|
|
611
|
-
|
|
612
|
-
|
|
613
|
-
|
|
614
|
-
-
|
|
615
|
-
|
|
631
|
+
The provider decides the configuration format, and the server rejects a
|
|
632
|
+
mismatch with 422 `invalid_provider_config_format`: `openai` needs the
|
|
633
|
+
`openai` block, `gemini` and `anthropic` need the `litellm` block (Gemini- or
|
|
634
|
+
Claude-specific settings under `litellm._overrides`), and the catalog
|
|
635
|
+
providers (`mistral`, `xai`, `huggingface`, `deepseek`, `cohere`, `groq`) need the
|
|
636
|
+
generic flat format — `model` and `modelOptions` directly on this object,
|
|
637
|
+
never an `openai` or `litellm` block.
|
|
638
|
+
- |-
|
|
639
|
+
Catalog providers are BYOK-only at run time: the agent saves without a
|
|
640
|
+
`_connectionId`, but its first run fails with 422 `invalid_request` ("No API
|
|
641
|
+
key configured for provider …"). Always attach an HTTP connection to the
|
|
642
|
+
vendor's API (token auth) when the provider is a catalog provider.
|
|
643
|
+
model:
|
|
644
|
+
type: string
|
|
645
|
+
description: |-
|
|
646
|
+
Model identifier for a catalog provider (generic flat format) — for example
|
|
647
|
+
`grok-4.6`. Built-in providers set the model inside their own block
|
|
648
|
+
(`openai.model`, `litellm.model`) instead.
|
|
649
|
+
x-celigo-ai-guidance:
|
|
650
|
+
- |-
|
|
651
|
+
Required for every catalog provider — omitting it or sending an empty string
|
|
652
|
+
fails the save with 422 `missing_required_field`.
|
|
653
|
+
- |-
|
|
654
|
+
Without a `_connectionId` on the parent import the model must be one the
|
|
655
|
+
catalog lists for the provider: an unknown id fails the save with 422
|
|
656
|
+
`model_not_found`, a listed-but-disabled id with 422 `model_disabled`. With a
|
|
657
|
+
`_connectionId` (BYOK) any model string saves — the server does not check it
|
|
658
|
+
against the catalog or the vendor.
|
|
659
|
+
modelOptions:
|
|
660
|
+
type: object
|
|
661
|
+
description: |-
|
|
662
|
+
Model tuning for a catalog provider (generic flat format). Which keys, values, and
|
|
663
|
+
ranges are accepted is defined per catalog model and validated on save when the
|
|
664
|
+
model is a catalog entry.
|
|
665
|
+
properties:
|
|
666
|
+
maxOutputTokens:
|
|
667
|
+
type: number
|
|
668
|
+
description: Maximum number of tokens the model may generate.
|
|
669
|
+
temperature:
|
|
670
|
+
type: number
|
|
671
|
+
description: Sampling temperature.
|
|
672
|
+
topP:
|
|
673
|
+
type: number
|
|
674
|
+
description: Nucleus sampling threshold.
|
|
675
|
+
reasoning_effort:
|
|
676
|
+
type: string
|
|
677
|
+
enum:
|
|
678
|
+
- low
|
|
679
|
+
- medium
|
|
680
|
+
- high
|
|
681
|
+
- xhigh
|
|
682
|
+
x-enumDescriptions:
|
|
683
|
+
low: Minimal reasoning before answering.
|
|
684
|
+
medium: Moderate reasoning.
|
|
685
|
+
high: Extensive reasoning.
|
|
686
|
+
xhigh: Maximum reasoning.
|
|
687
|
+
description: Reasoning depth for models that expose it.
|
|
688
|
+
x-celigo-ai-guidance:
|
|
689
|
+
- |-
|
|
690
|
+
When the model is a catalog entry, the server validates this object against
|
|
691
|
+
that model's schema and fails the save with 422 `invalid_model_options`
|
|
692
|
+
naming the key: unknown keys are rejected, `reasoning_effort` must be one of
|
|
693
|
+
the enum values where the model accepts it at all (xAI's `grok-4.6` does;
|
|
694
|
+
DeepSeek's `deepseek-chat` rejects the key), and numeric bounds apply —
|
|
695
|
+
on `grok-4.6`, `temperature` ≤ 2, `topP` ≤ 1, `maxOutputTokens` ≥ 100.
|
|
696
|
+
- |-
|
|
697
|
+
With a `_connectionId` and a model the catalog does not list, the object is
|
|
698
|
+
passed through unvalidated — send only options the vendor's API accepts.
|
|
699
|
+
instructions:
|
|
700
|
+
type: string
|
|
701
|
+
description: |-
|
|
702
|
+
System prompt for a catalog provider (generic flat format). Built-in providers
|
|
703
|
+
set it inside their own block (`openai.instructions`,
|
|
704
|
+
`litellm._overrides.anthropic.systemInstruction`).
|
|
705
|
+
output:
|
|
706
|
+
type: object
|
|
707
|
+
description: |-
|
|
708
|
+
Output format for a catalog provider (generic flat format). Built-in providers
|
|
709
|
+
configure it inside their own block (`openai.output`, `litellm.responseFormat`).
|
|
710
|
+
properties:
|
|
711
|
+
format:
|
|
712
|
+
type: object
|
|
713
|
+
description: Controls the structure of the model's output.
|
|
714
|
+
properties:
|
|
715
|
+
type:
|
|
716
|
+
type: string
|
|
717
|
+
enum:
|
|
718
|
+
- text
|
|
719
|
+
- json_schema
|
|
720
|
+
x-enumDescriptions:
|
|
721
|
+
text: Free-form text.
|
|
722
|
+
json_schema: Structured JSON conforming to `jsonSchema`.
|
|
723
|
+
description: Output type.
|
|
724
|
+
x-celigo-ai-guidance:
|
|
725
|
+
- |-
|
|
726
|
+
`blob` is not accepted here — the save fails with 422
|
|
727
|
+
`unsupported_output_format` naming the model (verified on `grok-4.6`).
|
|
728
|
+
name:
|
|
729
|
+
type: string
|
|
730
|
+
description: Name of the JSON schema, for `json_schema` output.
|
|
731
|
+
strict:
|
|
732
|
+
type: boolean
|
|
733
|
+
description: When true, the model must conform exactly to `jsonSchema`.
|
|
734
|
+
jsonSchema:
|
|
735
|
+
type: object
|
|
736
|
+
additionalProperties: true
|
|
737
|
+
description: JSON Schema the structured output must conform to, for `json_schema` output.
|
|
738
|
+
verbose:
|
|
739
|
+
type: string
|
|
740
|
+
description: Level of detail in the model's response, for models that expose it.
|
|
741
|
+
tools:
|
|
742
|
+
type: array
|
|
743
|
+
description: |-
|
|
744
|
+
Tools available to a catalog-provider agent (generic flat format). Entries mirror
|
|
745
|
+
the built-in providers' tool entries: a Celigo Tool by `_toolId`, an MCP server by
|
|
746
|
+
`_mcpConnectionId`, or a vendor web-search tool.
|
|
747
|
+
items:
|
|
748
|
+
type: object
|
|
749
|
+
properties:
|
|
750
|
+
type:
|
|
751
|
+
type: string
|
|
752
|
+
enum:
|
|
753
|
+
- tool
|
|
754
|
+
- mcp
|
|
755
|
+
- web_search
|
|
756
|
+
x-enumDescriptions:
|
|
757
|
+
tool: Reference a reusable Celigo Tool resource.
|
|
758
|
+
mcp: Connect to an MCP server for additional external tools.
|
|
759
|
+
web_search: Search the web for real-time information, where the vendor's model supports it.
|
|
760
|
+
description: Type of tool entry.
|
|
761
|
+
tool:
|
|
762
|
+
type: object
|
|
763
|
+
description: Reference to a Celigo Tool resource, used when type is "tool".
|
|
764
|
+
properties:
|
|
765
|
+
_toolId:
|
|
766
|
+
type: string
|
|
767
|
+
format: objectId
|
|
768
|
+
x-celigo-refModel: tools
|
|
769
|
+
description: The Celigo Tool to call.
|
|
770
|
+
mcp:
|
|
771
|
+
type: object
|
|
772
|
+
description: MCP server connection, used when type is "mcp".
|
|
773
|
+
properties:
|
|
774
|
+
_mcpConnectionId:
|
|
775
|
+
type: string
|
|
776
|
+
format: objectId
|
|
777
|
+
x-celigo-refModel: connections
|
|
778
|
+
description: Connection to the MCP server.
|
|
779
|
+
allowedTools:
|
|
780
|
+
type: array
|
|
781
|
+
description: |-
|
|
782
|
+
Specific tools to allow from the MCP server (all if omitted). Each entry
|
|
783
|
+
is either a plain tool name or an object carrying display metadata.
|
|
784
|
+
items:
|
|
785
|
+
type:
|
|
786
|
+
- string
|
|
787
|
+
- object
|
|
788
|
+
config:
|
|
789
|
+
type: object
|
|
790
|
+
additionalProperties: true
|
|
791
|
+
description: Tool-specific options for vendor-native tools (for example web-search settings).
|
|
792
|
+
x-celigo-ai-guidance:
|
|
793
|
+
- |-
|
|
794
|
+
References are checked on save: an unknown `_toolId` or `_mcpConnectionId`
|
|
795
|
+
fails with 422 `invalid_ref` naming the array index.
|
|
796
|
+
prompts:
|
|
797
|
+
type: array
|
|
798
|
+
description: |-
|
|
799
|
+
MCP prompt entries available to a catalog-provider agent (generic flat format).
|
|
800
|
+
Each entry references one MCP connection and the prompt names allowed from it.
|
|
801
|
+
items:
|
|
802
|
+
type: object
|
|
803
|
+
properties:
|
|
804
|
+
type:
|
|
805
|
+
type: string
|
|
806
|
+
enum:
|
|
807
|
+
- mcp
|
|
808
|
+
x-enumDescriptions:
|
|
809
|
+
mcp: Reference prompts exposed by an MCP server.
|
|
810
|
+
description: Type of prompt entry. Always "mcp".
|
|
811
|
+
mcp:
|
|
812
|
+
type: object
|
|
813
|
+
properties:
|
|
814
|
+
_mcpConnectionId:
|
|
815
|
+
type: string
|
|
816
|
+
format: objectId
|
|
817
|
+
x-celigo-refModel: connections
|
|
818
|
+
description: Connection to the MCP server that exposes the prompts.
|
|
819
|
+
allowedPrompts:
|
|
820
|
+
type: array
|
|
821
|
+
description: Prompt names the agent may fetch from the server.
|
|
822
|
+
items:
|
|
823
|
+
type: string
|
|
824
|
+
x-celigo-ai-guidance:
|
|
825
|
+
- |-
|
|
826
|
+
An unknown `_mcpConnectionId` fails the save with 422 `invalid_ref` naming
|
|
827
|
+
the array index.
|
|
828
|
+
resources:
|
|
829
|
+
$ref: '#/components/schemas/McpResources'
|
|
616
830
|
openai:
|
|
617
831
|
type: object
|
|
618
832
|
description: |
|
|
@@ -643,10 +857,11 @@ components:
|
|
|
643
857
|
x-celigo-ai-guidance:
|
|
644
858
|
- |-
|
|
645
859
|
With platform-managed credentials (no `_connectionId` on the parent
|
|
646
|
-
import), the model must be
|
|
647
|
-
unlisted model
|
|
648
|
-
|
|
649
|
-
|
|
860
|
+
import), the model must be one the platform hosts. The save does not
|
|
861
|
+
check this — an unlisted model saves, and the agent's first run fails
|
|
862
|
+
with 422 `invalid_model`, whose message lists the hosted models. A
|
|
863
|
+
BYOK connection bypasses the hosted list and the model string is
|
|
864
|
+
passed through to the provider as-is.
|
|
650
865
|
reasoning:
|
|
651
866
|
type: object
|
|
652
867
|
description: Controls depth of reasoning for complex tasks.
|
|
@@ -1049,6 +1264,13 @@ components:
|
|
|
1049
1264
|
- claude-sonnet-4-6
|
|
1050
1265
|
- claude-opus-4-7
|
|
1051
1266
|
x-celigo-ui-override: Required by the AI agent configuration form. Encoded here to mirror the form so builders produce connectable configurations.
|
|
1267
|
+
x-celigo-ai-guidance:
|
|
1268
|
+
- |-
|
|
1269
|
+
With platform-managed credentials (no `_connectionId` on the parent
|
|
1270
|
+
import), the model must be one the platform hosts for the provider —
|
|
1271
|
+
Gemini and Anthropic Claude both have a hosted list. The save does not
|
|
1272
|
+
check this — an unlisted model saves, and the agent's first run fails
|
|
1273
|
+
with 422 `invalid_model`, whose message lists the hosted models.
|
|
1052
1274
|
temperature:
|
|
1053
1275
|
type: number
|
|
1054
1276
|
minimum: 0
|
|
@@ -1703,6 +1925,22 @@ components:
|
|
|
1703
1925
|
litellm:
|
|
1704
1926
|
required:
|
|
1705
1927
|
- model
|
|
1928
|
+
else:
|
|
1929
|
+
if:
|
|
1930
|
+
properties:
|
|
1931
|
+
provider:
|
|
1932
|
+
enum:
|
|
1933
|
+
- mistral
|
|
1934
|
+
- xai
|
|
1935
|
+
- huggingface
|
|
1936
|
+
- deepseek
|
|
1937
|
+
- cohere
|
|
1938
|
+
- groq
|
|
1939
|
+
required:
|
|
1940
|
+
- provider
|
|
1941
|
+
then:
|
|
1942
|
+
required:
|
|
1943
|
+
- model
|
|
1706
1944
|
GuardrailConfig:
|
|
1707
1945
|
type: object
|
|
1708
1946
|
description: |-
|
|
@@ -2355,7 +2593,7 @@ components:
|
|
|
2355
2593
|
type: string
|
|
2356
2594
|
format: objectId
|
|
2357
2595
|
x-celigo-refModel: connections
|
|
2358
|
-
description: BYOK connection backing an `ai_agent` guardrail. Absent when platform-managed credentials
|
|
2596
|
+
description: BYOK connection backing an `ai_agent` guardrail. Absent when a built-in provider runs on platform-managed credentials.
|
|
2359
2597
|
examples:
|
|
2360
2598
|
- 68ae48d6b5f755d2dd38c294
|
|
2361
2599
|
_sourceId:
|
|
@@ -2489,7 +2727,12 @@ components:
|
|
|
2489
2727
|
type: string
|
|
2490
2728
|
format: objectId
|
|
2491
2729
|
x-celigo-refModel: connections
|
|
2492
|
-
description:
|
|
2730
|
+
description: |-
|
|
2731
|
+
Connection for BYOK (bring your own key) — only meaningful for `ai_agent` type
|
|
2732
|
+
guardrails. Optional for the built-in providers (`openai`, `gemini`, `anthropic`),
|
|
2733
|
+
which run on platform-managed credentials when it is omitted; catalog providers
|
|
2734
|
+
(`mistral`, `xai`, `huggingface`, `deepseek`, `cohere`, `groq`) have no platform-managed
|
|
2735
|
+
credentials and need it to run.
|
|
2493
2736
|
examples:
|
|
2494
2737
|
- 68ae48d6b5f755d2dd38c294
|
|
2495
2738
|
adaptorType:
|