copilotkit 4.9.1 → 4.9.4

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.
Files changed (45) hide show
  1. package/cli-build-info.json +8 -8
  2. package/index.js +151 -21
  3. package/onboarding/index.json +11 -0
  4. package/onboarding/prompts/authenticate/start.md +83 -32
  5. package/onboarding/prompts/credentials/finalize-plan.md +51 -33
  6. package/onboarding/prompts/credentials/plan.md +37 -25
  7. package/onboarding/prompts/fallback/best-effort.md +22 -11
  8. package/onboarding/prompts/framework/ag2.md +4 -15
  9. package/onboarding/prompts/framework/agno.md +4 -15
  10. package/onboarding/prompts/framework/built-in.md +4 -14
  11. package/onboarding/prompts/framework/claude-sdk-python.md +4 -14
  12. package/onboarding/prompts/framework/claude-sdk-typescript.md +4 -14
  13. package/onboarding/prompts/framework/crewai-flows.md +11 -23
  14. package/onboarding/prompts/framework/deep-agents.md +4 -15
  15. package/onboarding/prompts/framework/google-adk.md +3 -15
  16. package/onboarding/prompts/framework/langgraph-fastapi.md +3 -15
  17. package/onboarding/prompts/framework/langgraph-python.md +3 -15
  18. package/onboarding/prompts/framework/langgraph-typescript.md +3 -15
  19. package/onboarding/prompts/framework/llamaindex.md +4 -15
  20. package/onboarding/prompts/framework/mastra.md +3 -15
  21. package/onboarding/prompts/framework/ms-agent-dotnet.md +18 -49
  22. package/onboarding/prompts/framework/ms-agent-harness-dotnet.md +20 -30
  23. package/onboarding/prompts/framework/ms-agent-python.md +3 -15
  24. package/onboarding/prompts/framework/pydantic-ai.md +4 -14
  25. package/onboarding/prompts/framework/strands-python.md +4 -14
  26. package/onboarding/prompts/framework/strands-typescript.md +4 -14
  27. package/onboarding/prompts/frontend/angular.md +13 -18
  28. package/onboarding/prompts/frontend/nextjs.md +20 -15
  29. package/onboarding/prompts/frontend/plan.md +9 -14
  30. package/onboarding/prompts/frontend/react-native.md +7 -15
  31. package/onboarding/prompts/frontend/react-spa.md +5 -15
  32. package/onboarding/prompts/frontend/vue.md +8 -15
  33. package/onboarding/prompts/implementation/build-and-validate.md +22 -13
  34. package/onboarding/prompts/proof/complete.md +32 -5
  35. package/onboarding/prompts/proof/oss-baseline.md +3 -3
  36. package/onboarding/prompts/proof/round-trip.md +197 -72
  37. package/onboarding/prompts/starter/clone.md +67 -0
  38. package/onboarding/prompts/subagent/create-plan.md +26 -15
  39. package/onboarding/prompts/subagent/implement-and-validate.md +19 -10
  40. package/onboarding/prompts/subagent/inspect-repository.md +19 -7
  41. package/onboarding/prompts/subagent/prove-oss-baseline.md +1 -1
  42. package/onboarding/prompts/subagent/prove-round-trip.md +176 -68
  43. package/onboarding/prompts/unsupported/no-validated-path.md +2 -2
  44. package/package.json +1 -1
  45. package/release/release-tool.js +15 -1
@@ -24,7 +24,7 @@ starts. Choose one starting state: empty, agent-only, frontend-only, both, both-
24
24
  Run this command with the exact selected slugs:
25
25
 
26
26
  ```text
27
- npx --yes copilotkit@4.9.1 onboard classify --starting-state <starting-state> --agent-framework <agent-framework> --frontend <frontend>
27
+ npx --yes copilotkit@4.9.4 onboard classify --starting-state <starting-state> --agent-framework <agent-framework> --frontend <frontend>
28
28
  ```
29
29
 
30
30
  Do not continue if a value is refused. Fix the value from the choices that the earlier
@@ -35,12 +35,10 @@ prompts gave, then run the command again.
35
35
  Use the target app or runtime directory from the repository findings. If a nested app owns
36
36
  the runtime `.env` file, do not default to the repository root.
37
37
 
38
- If valid project fields and a non-empty `INTELLIGENCE_API_KEY` already exist, ask whether to
39
- keep that selection. Recommend keeping it to avoid minting another key. Ask the developer to
40
- confirm that the prior project selection had no key-provision warning.
41
-
42
- Reused project credentials remain unverified until an authenticated Intelligence call
43
- succeeds. Record that state in the plan.
38
+ If valid project fields and a non-empty `INTELLIGENCE_API_KEY` already exist, use this
39
+ rule: Reuse that project without another question. Do not mint another key. The round-trip
40
+ proof later runs the authenticated Intelligence checks. Record the reused project
41
+ credentials as unverified in the plan until those checks pass.
44
42
 
45
43
  If the developer needs a project or another project, do the selection yourself from the
46
44
  target directory. Do not send the developer to another terminal.
@@ -48,7 +46,7 @@ target directory. Do not send the developer to another terminal.
48
46
  Read the choices first:
49
47
 
50
48
  ```text
51
- npx --yes copilotkit@4.9.1 project list --json
49
+ npx --yes copilotkit@4.9.4 project list --json
52
50
  ```
53
51
 
54
52
  With `--json` the payload is the only thing on standard output, so it is safe to parse.
@@ -59,7 +57,7 @@ developer asks.
59
57
  If the developer searches, run:
60
58
 
61
59
  ```text
62
- npx --yes copilotkit@4.9.1 project list --search <query> --json
60
+ npx --yes copilotkit@4.9.4 project list --search <query> --json
63
61
  ```
64
62
 
65
63
  If the developer asks for all projects, show the result of the first list command. Ask
@@ -69,7 +67,7 @@ frontend, framework, model, credential, or plan-approval question.
69
67
  Then record their answer, from the target directory:
70
68
 
71
69
  ```text
72
- npx --yes copilotkit@4.9.1 project select --project <slug-or-id> --json
70
+ npx --yes copilotkit@4.9.4 project select --project <slug-or-id> --json
73
71
  ```
74
72
 
75
73
  Use `--create <name> --json` instead when they want a new project. The two flags cannot be
@@ -123,28 +121,37 @@ it. The credential reaches the platform only when the runtime is constructed wit
123
121
  Intelligence client, and a runtime built without one compiles, serves, answers in a
124
122
  browser, and never touches the platform.
125
123
 
126
- Add this page to the selected documentation URLs for both subagents:
124
+ Add these pages to the selected documentation URLs for the planning, implementation, and
125
+ proof subagents:
127
126
 
128
127
  - https://docs.copilotkit.ai/premium/connect-your-runtime.md
128
+ - https://docs.copilotkit.ai/backend/runtime-endpoints.md
129
+ - https://docs.copilotkit.ai/premium/managed-intelligence-platform.md
129
130
 
130
- Fetch it together with the pages already selected rather than on its own.
131
+ Fetch them together with the pages already selected rather than on their own.
131
132
 
132
133
  Spawn one planning subagent. Give it the full text of the planning brief at the end of this
133
134
  prompt, the repository findings, selected framework, frontend, model, credential variable
134
- names, and selected documentation URLs. Wait for the subagent to finish.
135
+ names, selected documentation URLs, and documentation policy. Wait for the subagent to
136
+ finish.
135
137
 
136
138
  Make sure that the plan preserves each part that already exists. The plan must name the
137
139
  credential variables, the Intelligence runtime wiring, the application the project asks
138
- for, implementation steps, validation steps, and proof steps. It must not contain secret
139
- values.
140
+ for, implementation steps, validation steps, and proof steps. Where the plan upgrades the
141
+ CopilotKit dependencies, it must name the versions it moves and the revert it falls back
142
+ to. It must not contain secret values.
143
+
144
+ Show only the selected framework, frontend, model, planned file changes, the CopilotKit dependency upgrade and the versions it moves, validation commands, and proof steps.
140
145
 
141
- Show only the selected framework, frontend, model, planned file changes, validation commands, and proof steps.
146
+ An upgrade to a working install is the developer's call, so it belongs in the plan they
147
+ approve rather than in the implementation that follows. Do not upgrade a dependency the
148
+ developer did not approve.
142
149
 
143
150
  If the developer approves the plan, run
144
- `npx --yes copilotkit@4.9.1 onboard read implementation/build-and-validate`.
151
+ `npx --yes copilotkit@4.9.4 onboard read implementation/build-and-validate`.
145
152
 
146
153
  If no exact supported path or documentation URL exists, run
147
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
154
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.
148
155
 
149
156
  ## Planning subagent brief
150
157
 
@@ -155,20 +162,13 @@ Everything below the rule is the subagent's prompt. Give it verbatim.
155
162
  # Create the onboarding plan
156
163
 
157
164
  Use the selected framework, frontend, model, repository findings, and documentation URLs
158
- from the main coding agent. Fetch every selected URL in one step rather than one after
159
- another. Do not use remembered CopilotKit instructions.
160
-
161
- A fetch tool that refuses a URL, or fails to reach it, reports a limit of the tool and
162
- not a fact about the page. Retrieve the same URL a second way before you judge it. Run
163
- `curl -fsSL <url>`, or read the same page without the `.md` suffix. Report a
164
- documentation gap only after a second method also fails.
165
+ from the main coding agent. Follow the documentation policy it gives you.
165
166
 
166
167
  Name the application the project asks for, and plan that application. Each documentation
167
168
  page teaches through one worked example, and that example carries a domain of its own.
168
- The domain belongs to the page. The project's purpose from the repository findings
169
- decides what gets built. Where the two differ, the project wins. Where the repository
170
- never states a purpose, say so in the plan and ask the developer before you adopt the
171
- domain of an example.
169
+ The domain belongs to the page. Use the purpose from the repository findings, or the
170
+ developer outcome if the repository purpose was unproved. Do not ask for it again. Where
171
+ the documentation example differs from that purpose, the project wins.
172
172
 
173
173
  Read only inside the project directory. Answer an API question from the documentation
174
174
  URLs you were given, not from another checkout on this machine.
@@ -185,10 +185,28 @@ combined. Take the constructor from the connect-your-runtime page. Where a frame
185
185
  quickstart shows a `runner` option instead, the connect-your-runtime page wins.
186
186
 
187
187
  When the starting state is `both-oss`, use its recorded live baseline evidence. Preserve
188
- the working agent, frontend, CopilotKit integration, and OSS behavior. Plan only the
189
- project selection, Intelligence runtime configuration, and authenticated proof needed for
190
- the conversion. Preserve the existing persistence and user-visible request. Do not
191
- rebuild a path that already works.
188
+ the working agent, frontend, CopilotKit integration, and OSS behavior. Plan the project
189
+ selection, the CopilotKit dependency upgrade below, the Intelligence runtime
190
+ configuration, and the authenticated proof needed for the conversion. Preserve the
191
+ existing persistence and user-visible request. Do not rebuild a path that already works.
192
+
193
+ Preserving the OSS baseline preserves the application, not its CopilotKit dependency
194
+ versions. An install that predates the managed platform defaults carries a bundled
195
+ reference naming hosts that route nothing, and it requires the platform URLs the same
196
+ documentation calls optional. A run that reads that reference configures the runtime
197
+ against a dead host, and the failure arrives as an empty-body 404 with no cause named.
198
+
199
+ If any `@copilotkit/*` dependency is below 1.64.0, plan to upgrade every `@copilotkit/*`
200
+ dependency to its latest published version. Packages that share a version line must end
201
+ on the same version. Never force a package onto a version line it does not publish on.
202
+ Resolve each latest version at run time rather than from a remembered version number.
203
+ Where every `@copilotkit/*` dependency already meets that floor, plan no dependency
204
+ change.
205
+
206
+ Plan the upgrade as its own step before the Intelligence runtime wiring, and plan to
207
+ re-run the recorded baseline checks immediately after it. Name the revert: restore the
208
+ manifest and lockfile to their recorded state and stop, rather than wiring Intelligence
209
+ onto a baseline the upgrade broke.
192
210
 
193
211
  Where the SDK requires an application-level value that the repository cannot supply, such
194
212
  as an end-user identity for threads, use one clearly marked local placeholder and name what
@@ -4,11 +4,23 @@ Use the repository findings to select the agent framework. Ask the developer onl
4
4
  choices that the repository does not show. Do not change application code in this phase.
5
5
  Do not read, show, store, or request a secret value.
6
6
 
7
+ Keep `INTELLIGENCE_API_KEY` separate from every model-vendor credential. Do not read,
8
+ show, store, or copy either credential.
9
+
10
+ This documentation policy applies to every framework and frontend route below. Deduplicate
11
+ the selected URL list and fetch each URL once. Fetch all selected URLs in one step. Do not
12
+ use remembered CopilotKit instructions. A fetch tool that refuses a URL, or fails to reach
13
+ it, reports a limit of the tool and not a fact about the page. Retrieve the same URL a
14
+ second way before you judge it. Run `curl -fsSL <url>`, or read the same page without the
15
+ `.md` suffix. Treat a page as unavailable only after a second method also fails. Give this
16
+ policy to every documentation subagent. A page that is still unavailable after the second
17
+ method does not support the selection.
18
+
7
19
  Use a gentle wizard. Ask one short question at a time. Do not combine separate choices in
8
20
  one question. Put the recommendation first, give its evidence in one sentence, and keep
9
- the alternatives short. When the framework is unknown, ask only whether to use the
10
- recommended framework. If the developer declines it, ask one follow-up question that lists
11
- the valid framework choices. Do not ask about the frontend, model, or project in this step.
21
+ the alternatives short. When the framework is unknown, use this rule: Ask one framework
22
+ question that lists the recommendation and valid choices. Do not ask a yes-or-no question
23
+ first. Do not ask about the frontend, model, or project in this step.
12
24
 
13
25
  These are the agent frameworks and their default vendors:
14
26
 
@@ -27,9 +39,9 @@ These are the agent frameworks and their default vendors:
27
39
  | LangGraph TypeScript | OpenAI |
28
40
  | LlamaIndex | OpenAI |
29
41
  | Mastra | OpenAI |
30
- | Microsoft Agent Framework (.NET) | GitHub Models |
42
+ | Microsoft Agent Framework (.NET) | OpenAI |
31
43
  | Microsoft Agent Framework (Python) | OpenAI |
32
- | MS Agent Harness (.NET) | GitHub Models |
44
+ | MS Agent Harness (.NET) | OpenAI |
33
45
  | Pydantic AI | OpenAI |
34
46
  | Strands Agents (Python) | OpenAI |
35
47
  | Strands Agents (TypeScript) | OpenAI |
@@ -51,26 +63,26 @@ another framework. Do not show the internal route.
51
63
 
52
64
  Use exactly one matching internal route:
53
65
 
54
- 1. AG2: `npx --yes copilotkit@4.9.1 onboard read framework/ag2`
55
- 2. Agno: `npx --yes copilotkit@4.9.1 onboard read framework/agno`
56
- 3. Built-in CopilotKit agent: `npx --yes copilotkit@4.9.1 onboard read framework/built-in`
57
- 4. Claude Agent SDK Python: `npx --yes copilotkit@4.9.1 onboard read framework/claude-sdk-python`
58
- 5. Claude Agent SDK TypeScript: `npx --yes copilotkit@4.9.1 onboard read framework/claude-sdk-typescript`
59
- 6. CrewAI Flows: `npx --yes copilotkit@4.9.1 onboard read framework/crewai-flows`
60
- 7. Deep Agents: `npx --yes copilotkit@4.9.1 onboard read framework/deep-agents`
61
- 8. LangGraph Python: `npx --yes copilotkit@4.9.1 onboard read framework/langgraph-python`
62
- 9. LangGraph FastAPI: `npx --yes copilotkit@4.9.1 onboard read framework/langgraph-fastapi`
63
- 10. LangGraph TypeScript: `npx --yes copilotkit@4.9.1 onboard read framework/langgraph-typescript`
64
- 11. LlamaIndex: `npx --yes copilotkit@4.9.1 onboard read framework/llamaindex`
65
- 12. ADK: `npx --yes copilotkit@4.9.1 onboard read framework/google-adk`
66
- 13. Microsoft Agent Framework Python: `npx --yes copilotkit@4.9.1 onboard read framework/ms-agent-python`
67
- 14. Microsoft Agent Framework .NET: `npx --yes copilotkit@4.9.1 onboard read framework/ms-agent-dotnet`
68
- 15. Mastra: `npx --yes copilotkit@4.9.1 onboard read framework/mastra`
69
- 16. MS Agent Harness .NET: `npx --yes copilotkit@4.9.1 onboard read framework/ms-agent-harness-dotnet`
70
- 17. Pydantic AI: `npx --yes copilotkit@4.9.1 onboard read framework/pydantic-ai`
71
- 18. Strands Agents Python: `npx --yes copilotkit@4.9.1 onboard read framework/strands-python`
72
- 19. Strands Agents TypeScript: `npx --yes copilotkit@4.9.1 onboard read framework/strands-typescript`
66
+ 1. AG2: `npx --yes copilotkit@4.9.4 onboard read framework/ag2`
67
+ 2. Agno: `npx --yes copilotkit@4.9.4 onboard read framework/agno`
68
+ 3. Built-in CopilotKit agent: `npx --yes copilotkit@4.9.4 onboard read framework/built-in`
69
+ 4. Claude Agent SDK Python: `npx --yes copilotkit@4.9.4 onboard read framework/claude-sdk-python`
70
+ 5. Claude Agent SDK TypeScript: `npx --yes copilotkit@4.9.4 onboard read framework/claude-sdk-typescript`
71
+ 6. CrewAI Flows: `npx --yes copilotkit@4.9.4 onboard read framework/crewai-flows`
72
+ 7. Deep Agents: `npx --yes copilotkit@4.9.4 onboard read framework/deep-agents`
73
+ 8. LangGraph Python: `npx --yes copilotkit@4.9.4 onboard read framework/langgraph-python`
74
+ 9. LangGraph FastAPI: `npx --yes copilotkit@4.9.4 onboard read framework/langgraph-fastapi`
75
+ 10. LangGraph TypeScript: `npx --yes copilotkit@4.9.4 onboard read framework/langgraph-typescript`
76
+ 11. LlamaIndex: `npx --yes copilotkit@4.9.4 onboard read framework/llamaindex`
77
+ 12. ADK: `npx --yes copilotkit@4.9.4 onboard read framework/google-adk`
78
+ 13. Microsoft Agent Framework Python: `npx --yes copilotkit@4.9.4 onboard read framework/ms-agent-python`
79
+ 14. Microsoft Agent Framework .NET: `npx --yes copilotkit@4.9.4 onboard read framework/ms-agent-dotnet`
80
+ 15. Mastra: `npx --yes copilotkit@4.9.4 onboard read framework/mastra`
81
+ 16. MS Agent Harness .NET: `npx --yes copilotkit@4.9.4 onboard read framework/ms-agent-harness-dotnet`
82
+ 17. Pydantic AI: `npx --yes copilotkit@4.9.4 onboard read framework/pydantic-ai`
83
+ 18. Strands Agents Python: `npx --yes copilotkit@4.9.4 onboard read framework/strands-python`
84
+ 19. Strands Agents TypeScript: `npx --yes copilotkit@4.9.4 onboard read framework/strands-typescript`
73
85
 
74
86
  If the project has an agent in another framework, or no listed framework fits, keep the
75
87
  developer's current agent and run
76
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
88
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.
@@ -1,19 +1,25 @@
1
1
  # Build from the best available official documentation
2
2
 
3
- Tell the implementation subagent which plan steps lack direct documentation. Give it the
4
- approved plan, repository findings, and all official documentation that you found.
3
+ Work only in the target project directory. Do not read, show, store, or request a secret
4
+ value.
5
+
6
+ Spawn one implementation subagent. Tell it which plan steps lack direct documentation.
7
+ Give it the approved plan, repository findings, target project directory, and all official
8
+ documentation that you fetched for the plan.
5
9
 
6
10
  Tell the subagent to preserve the existing project. It must not invent a CopilotKit API.
7
11
  It can combine compatible APIs from different official pages when one page omits the
8
12
  selected framework and frontend combination.
9
13
 
10
- Run the approved implementation and validation steps. Repair a failure in new work at most
11
- three times. Stop if the fix requires changes to the developer's existing agent or
12
- frontend.
14
+ Tell the implementation subagent to run the approved implementation and validation steps.
15
+ Send a failure in new work back to it at most three times. Stop if a fix requires changes
16
+ to the developer's existing agent or frontend.
13
17
 
14
- After validation passes, prove the runtime, agent round trip, and real frontend path. Use
15
- the same proof rules from the approved plan. Keep the development servers running after
16
- proof.
18
+ After validation passes, use this rule: Spawn one proof subagent. Give it the approved
19
+ plan, validation evidence, target project directory, all official documentation that you
20
+ fetched for the plan, and recorded browser or device control. Tell it to prove the runtime,
21
+ agent round trip, and real frontend path with the same proof rules from the approved plan.
22
+ Keep the development servers running after proof.
17
23
 
18
24
  Do not block core proof on CopilotKit Skills or MCP configuration. Try these tools after
19
25
  the application passes proof. Report each tool result separately.
@@ -21,16 +27,21 @@ the application passes proof. Report each tool result separately.
21
27
  Report the documentation gap and each assumption with the proof evidence. Do not claim
22
28
  that the selected documentation proved an inferred step.
23
29
 
24
- When the proof is complete, run `npx --yes copilotkit@4.9.1 onboard complete`, carrying the
30
+ When the proof is complete, run `npx --yes copilotkit@4.9.4 onboard complete`, carrying the
25
31
  surface-check result the proof subagent returned. Pass exactly one flag, matching this
26
32
  journey's surface:
27
33
 
28
34
  ```text
29
- npx --yes copilotkit@4.9.1 onboard complete --visual-check <outcome>
30
- npx --yes copilotkit@4.9.1 onboard complete --device-check <outcome>
35
+ npx --yes copilotkit@4.9.4 onboard complete --visual-check <outcome>
36
+ npx --yes copilotkit@4.9.4 onboard complete --device-check <outcome>
31
37
  ```
32
38
 
33
39
  `--visual-check` is for a web frontend and takes `performed`, `skipped-no-browser-tool`, or
34
40
  `failed`. `--device-check` is for React Native and takes `performed`, `skipped-no-device`,
35
41
  or `failed`. Anything but `performed` ends this run as blocked, and the command's output
36
42
  names the evidence that is missing. Report it that way.
43
+
44
+ If the proof passed and something after it still blocked this run, add
45
+ `--blocked-by <cause>` to the same command, with one of `hosted-dashboard`, `inspector`,
46
+ `plan-excluded-capability`, or `other`. It ends the run as blocked and names what the
47
+ blocker leaves unverified.
@@ -6,9 +6,6 @@ credential variables and setup steps.
6
6
 
7
7
  Name the default model credential `OPENAI_API_KEY`.
8
8
 
9
- Keep `INTELLIGENCE_API_KEY` separate from model-vendor credentials. Do not read, show,
10
- store, or request a secret value.
11
-
12
9
  ## Documentation
13
10
 
14
11
  - https://docs.copilotkit.ai/ag2/quickstart.md
@@ -16,8 +13,6 @@ store, or request a secret value.
16
13
  - https://docs.copilotkit.ai/ag2/generative-ui/a2ui/fixed-schema.md
17
14
  - https://docs.copilotkit.ai/ag2/readables.md
18
15
 
19
- Fetch all four pages in one step rather than one after another.
20
-
21
16
  The documentation scope is `ag2`. Use only the Python examples from these pages.
22
17
  Preserve the existing `AGUIStream` endpoint and its route. The quickstart uses a sample
23
18
  repository with `uv sync`. It does not name a dependency list for an existing project.
@@ -33,16 +28,10 @@ The Readables page includes the AG2 backend context path. Use its `ContextVariab
33
28
  framework's context setup.
34
29
 
35
30
  Record the selected framework, vendor, model, required credential variable names, these
36
- URLs, and the fixed-schema documentation gap. Do not use remembered CopilotKit
37
- instructions.
38
-
39
- A fetch tool that refuses a URL, or fails to reach it, reports a limit of the tool and
40
- not a fact about the page. Retrieve the same URL a second way before you judge it. Run
41
- `curl -fsSL <url>`, or read the same page without the `.md` suffix. Treat a page as
42
- unavailable only after a second method also fails.
31
+ URLs, and the fixed-schema documentation gap.
43
32
 
44
33
  If the pages support the selection, run
45
- `npx --yes copilotkit@4.9.1 onboard read frontend/plan`.
34
+ `npx --yes copilotkit@4.9.4 onboard read frontend/plan`.
46
35
 
47
- If a second method also fails, or a page that loads does not support the selection, run
48
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
36
+ If the documentation does not support the selection, run
37
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.
@@ -6,9 +6,6 @@ credential variables and setup steps.
6
6
 
7
7
  Name the default model credential `OPENAI_API_KEY`.
8
8
 
9
- Keep `INTELLIGENCE_API_KEY` separate from model-vendor credentials. Do not read, show,
10
- store, or request a secret value.
11
-
12
9
  ## Documentation
13
10
 
14
11
  - https://docs.copilotkit.ai/agno/quickstart.md
@@ -16,8 +13,6 @@ store, or request a secret value.
16
13
  - https://docs.copilotkit.ai/agno/generative-ui/a2ui/fixed-schema.md
17
14
  - https://docs.copilotkit.ai/agno/shared-state/agent-readonly.md
18
15
 
19
- Fetch all four pages in one step rather than one after another.
20
-
21
16
  The documentation scope is `agno`. Use only the Python agent examples from these pages.
22
17
  When the repository already has an Agno agent, use the quickstart's
23
18
  `Use an existing agent` path. Use `Start from scratch` only when the repository needs a
@@ -33,16 +28,10 @@ Do not invent either missing backend setup. If the requested work needs one of t
33
28
  setups and the project does not already have it, run the unsupported route below.
34
29
 
35
30
  Record the selected framework, vendor, model, required credential variable names, these
36
- URLs, the documented dependencies and adapter, and both documentation gaps. Do not use
37
- remembered CopilotKit instructions.
38
-
39
- A fetch tool that refuses a URL, or fails to reach it, reports a limit of the tool and
40
- not a fact about the page. Retrieve the same URL a second way before you judge it. Run
41
- `curl -fsSL <url>`, or read the same page without the `.md` suffix. Treat a page as
42
- unavailable only after a second method also fails.
31
+ URLs, the documented dependencies and adapter, and both documentation gaps.
43
32
 
44
33
  If the pages support the selection, run
45
- `npx --yes copilotkit@4.9.1 onboard read frontend/plan`.
34
+ `npx --yes copilotkit@4.9.4 onboard read frontend/plan`.
46
35
 
47
- If a second method also fails, or a page that loads does not support the selection, run
48
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
36
+ If the documentation does not support the selection, run
37
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.
@@ -6,9 +6,6 @@ its credential variables and setup steps.
6
6
 
7
7
  Name the default model credential `OPENAI_API_KEY`.
8
8
 
9
- Keep `INTELLIGENCE_API_KEY` separate from model-vendor credentials. Do not read, show,
10
- store, or request a secret value.
11
-
12
9
  The built-in agent runs in the CopilotKit Runtime process.
13
10
  Do not create or start a separate agent server.
14
11
  Keep the agent, model choice, tools, and runtime endpoint in the project's existing
@@ -23,21 +20,14 @@ runtime setup.
23
20
  - https://docs.copilotkit.ai/model-selection.md
24
21
  - https://docs.copilotkit.ai/generative-ui/tool-based.md
25
22
 
26
- Fetch all six pages in one step rather than one after another.
27
-
28
23
  These pages use the root documentation scope. Do not add a framework path segment to
29
24
  their URLs. Use the tool-based generative UI page for this built-in route.
30
25
 
31
26
  Record the selected framework, vendor, model, required credential variable names, and
32
- these URLs. Do not use remembered CopilotKit instructions.
33
-
34
- A fetch tool that refuses a URL, or fails to reach it, reports a limit of the tool and
35
- not a fact about the page. Retrieve the same URL a second way before you judge it. Run
36
- `curl -fsSL <url>`, or read the same page without the `.md` suffix. Treat a page as
37
- unavailable only after a second method also fails.
27
+ these URLs.
38
28
 
39
29
  If the pages support the selection, run
40
- `npx --yes copilotkit@4.9.1 onboard read frontend/plan`.
30
+ `npx --yes copilotkit@4.9.4 onboard read frontend/plan`.
41
31
 
42
- If a second method also fails, or a page that loads does not support the selection, run
43
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
32
+ If the documentation does not support the selection, run
33
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.
@@ -7,9 +7,6 @@ name its credential variables and setup steps.
7
7
  Name the model credential `ANTHROPIC_API_KEY`. The documented model variable is
8
8
  `ANTHROPIC_MODEL`.
9
9
 
10
- Keep `INTELLIGENCE_API_KEY` separate from model-vendor credentials. Do not read, show,
11
- store, or request a secret value.
12
-
13
10
  ## Documentation
14
11
 
15
12
  - https://docs.copilotkit.ai/claude-sdk-python/quickstart.md
@@ -17,8 +14,6 @@ store, or request a secret value.
17
14
  - https://docs.copilotkit.ai/claude-sdk-python/generative-ui/a2ui/fixed-schema.md
18
15
  - https://docs.copilotkit.ai/reference/v2/hooks/useAgentContext.md
19
16
 
20
- Fetch all four pages in one step rather than one after another.
21
-
22
17
  The documentation scope is `claude-sdk-python`. For an existing agent, install the exact
23
18
  documented backend set with
24
19
  `uv add claude-agent-sdk ag-ui-claude-sdk ag-ui-protocol anthropic fastapi uvicorn python-dotenv`.
@@ -35,15 +30,10 @@ page above documents the frontend hook only. Do not invent the missing agent-sid
35
30
  setup. If the requested work needs that setup, run the unsupported route below.
36
31
 
37
32
  Record the selected framework, vendor, model, required credential variable names, these
38
- URLs, and the context and A2UI limits. Do not use remembered CopilotKit instructions.
39
-
40
- A fetch tool that refuses a URL, or fails to reach it, reports a limit of the tool and
41
- not a fact about the page. Retrieve the same URL a second way before you judge it. Run
42
- `curl -fsSL <url>`, or read the same page without the `.md` suffix. Treat a page as
43
- unavailable only after a second method also fails.
33
+ URLs, and the context and A2UI limits.
44
34
 
45
35
  If the pages support the selection, run
46
- `npx --yes copilotkit@4.9.1 onboard read frontend/plan`.
36
+ `npx --yes copilotkit@4.9.4 onboard read frontend/plan`.
47
37
 
48
- If a second method also fails, or a page that loads does not support the selection, run
49
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
38
+ If the documentation does not support the selection, run
39
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.
@@ -7,9 +7,6 @@ name its credential variables and setup steps.
7
7
  Name the model credential `ANTHROPIC_API_KEY`. The documented model variable is
8
8
  `CLAUDE_MODEL`.
9
9
 
10
- Keep `INTELLIGENCE_API_KEY` separate from model-vendor credentials. Do not read, show,
11
- store, or request a secret value.
12
-
13
10
  ## Documentation
14
11
 
15
12
  - https://docs.copilotkit.ai/claude-sdk-typescript/quickstart.md
@@ -17,8 +14,6 @@ store, or request a secret value.
17
14
  - https://docs.copilotkit.ai/claude-sdk-typescript/generative-ui/a2ui/fixed-schema.md
18
15
  - https://docs.copilotkit.ai/reference/v2/hooks/useAgentContext.md
19
16
 
20
- Fetch all four pages in one step rather than one after another.
21
-
22
17
  The documentation scope is `claude-sdk-typescript`. For an existing agent, install the
23
18
  documented packages, including `@anthropic-ai/claude-agent-sdk@^0.2.58`,
24
19
  `@anthropic-ai/sdk`, `@ag-ui/claude-agent-sdk`, `@ag-ui/core`, `@ag-ui/encoder`,
@@ -39,15 +34,10 @@ pinned adapter source before using this path. If the package no longer handles
39
34
  `input.context`, run the unsupported route below.
40
35
 
41
36
  Record the selected framework, vendor, model, required credential variable names, these
42
- URLs, and the adapter and context limits. Do not use remembered CopilotKit instructions.
43
-
44
- A fetch tool that refuses a URL, or fails to reach it, reports a limit of the tool and
45
- not a fact about the page. Retrieve the same URL a second way before you judge it. Run
46
- `curl -fsSL <url>`, or read the same page without the `.md` suffix. Treat a page as
47
- unavailable only after a second method also fails.
37
+ URLs, and the adapter and context limits.
48
38
 
49
39
  If the pages support the selection, run
50
- `npx --yes copilotkit@4.9.1 onboard read frontend/plan`.
40
+ `npx --yes copilotkit@4.9.4 onboard read frontend/plan`.
51
41
 
52
- If a second method also fails, or a page that loads does not support the selection, run
53
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
42
+ If the documentation does not support the selection, run
43
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.
@@ -1,12 +1,9 @@
1
1
  # Configure CrewAI Flows
2
2
 
3
- Preserve the project's existing CrewAI Flow, model, and deployment. The selected pages do
4
- not name a default model vendor or model credential variable. Do not choose a vendor or
5
- credential name for the developer. Ask only for a choice that the repository does not
6
- show and the documentation supports.
7
-
8
- Keep `INTELLIGENCE_API_KEY` separate from model-vendor credentials. Do not read, show,
9
- store, or request a secret value.
3
+ Preserve the project's existing CrewAI Flow, provider, model, and deployment. The selected
4
+ pages do not name a default model vendor or model credential variable. If neither the
5
+ repository nor these pages prove a provider, run the unsupported route below. Do not
6
+ choose a vendor or credential name for the developer.
10
7
 
11
8
  ## Documentation
12
9
 
@@ -15,16 +12,13 @@ store, or request a secret value.
15
12
  - https://docs.copilotkit.ai/crewai-crews/generative-ui/a2ui/fixed-schema.md
16
13
  - https://docs.copilotkit.ai/crewai-crews/shared-state/in-app-agent-read.md
17
14
 
18
- Fetch all four pages in one step rather than one after another.
19
-
20
15
  The public onboarding route ID is `crewai-flows`. The canonical documentation scope is
21
16
  `crewai-crews`. The quickstart and feature pages also contain older `/crewai-flows/`
22
17
  links. Use the canonical Markdown URLs above and keep the project's CrewAI Flow name.
23
18
 
24
19
  The quickstart offers `Use the CopilotKit CLI (NextJS only)` and `Code along`. It assumes
25
- an existing Flow. Preserve the vendor, model, and credential variable that the project
26
- already uses. The new quickstart example uses GitHub Models. An existing OpenAI Flow can
27
- keep OpenAI and `OPENAI_API_KEY`. Do not change vendors to match the example.
20
+ an existing Flow. Preserve the provider, model, and credential variable that the project
21
+ already uses. Do not change the provider to match an example.
28
22
 
29
23
  The current Python base adapter is `ag-ui-crewai==0.3.0`. Add the endpoint to the
30
24
  project's FastAPI app with
@@ -40,17 +34,11 @@ The shared-state page is not an agent-app-context guide. Do not use it to invent
40
34
  `useAgentContext` backend setup. If the requested work needs app context rather than the
41
35
  documented shared-state path, run the unsupported route below.
42
36
 
43
- Record the selected framework, project-defined vendor and model, repository-proven
44
- credential variable names, these URLs, the adapter, and both documentation limits. Do not
45
- use remembered CopilotKit instructions.
46
-
47
- A fetch tool that refuses a URL, or fails to reach it, reports a limit of the tool and
48
- not a fact about the page. Retrieve the same URL a second way before you judge it. Run
49
- `curl -fsSL <url>`, or read the same page without the `.md` suffix. Treat a page as
50
- unavailable only after a second method also fails.
37
+ Record the selected framework, project-defined provider and model, repository-proven
38
+ credential variable names, these URLs, the adapter, and both documentation limits.
51
39
 
52
40
  If the pages support the selection, run
53
- `npx --yes copilotkit@4.9.1 onboard read frontend/plan`.
41
+ `npx --yes copilotkit@4.9.4 onboard read frontend/plan`.
54
42
 
55
- If a second method also fails, or a page that loads does not support the selection, run
56
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
43
+ If the documentation does not support the selection, run
44
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.
@@ -6,9 +6,6 @@ page must name its credential variables and setup steps.
6
6
 
7
7
  Name the default model credential `OPENAI_API_KEY`.
8
8
 
9
- Keep `INTELLIGENCE_API_KEY` separate from model-vendor credentials. Do not read, show,
10
- store, or request a secret value.
11
-
12
9
  ## Documentation
13
10
 
14
11
  - https://docs.copilotkit.ai/deepagents/quickstart.md
@@ -16,8 +13,6 @@ store, or request a secret value.
16
13
  - https://docs.copilotkit.ai/deepagents/generative-ui/a2ui/fixed-schema.md
17
14
  - https://docs.copilotkit.ai/deepagents/shared-state/agent-readonly.md
18
15
 
19
- Fetch all four pages in one step rather than one after another.
20
-
21
16
  The documentation scope is `deepagents`. Do not use `deep-agents` in a documentation URL.
22
17
  The quickstart has Python and TypeScript tabs. Preserve the project's current language.
23
18
  Preserve the project's existing deployment path. Choose the same `Deep Agent` or
@@ -30,16 +25,10 @@ agent-app-context setup. If the requested work needs either missing setup, run t
30
25
  unsupported route below.
31
26
 
32
27
  Record the selected framework, language, vendor, model, required credential variable
33
- names, these URLs, and each documentation gap. Do not use remembered CopilotKit
34
- instructions.
35
-
36
- A fetch tool that refuses a URL, or fails to reach it, reports a limit of the tool and
37
- not a fact about the page. Retrieve the same URL a second way before you judge it. Run
38
- `curl -fsSL <url>`, or read the same page without the `.md` suffix. Treat a page as
39
- unavailable only after a second method also fails.
28
+ names, these URLs, and each documentation gap.
40
29
 
41
30
  If the pages support the selection, run
42
- `npx --yes copilotkit@4.9.1 onboard read frontend/plan`.
31
+ `npx --yes copilotkit@4.9.4 onboard read frontend/plan`.
43
32
 
44
- If a second method also fails, or a page that loads does not support the selection, run
45
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
33
+ If the documentation does not support the selection, run
34
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.
@@ -8,9 +8,6 @@ Name the model credential `GOOGLE_API_KEY`. This framework accepts more than one
8
8
  spelling, and a scaffold that reaches for a different one looks finished and reads no
9
9
  key.
10
10
 
11
- Keep `INTELLIGENCE_API_KEY` separate from model-vendor credentials. Do not read, show,
12
- store, or request a secret value.
13
-
14
11
  ## Documentation
15
12
 
16
13
  - https://docs.copilotkit.ai/google-adk/quickstart.md
@@ -18,9 +15,6 @@ store, or request a secret value.
18
15
  - https://docs.copilotkit.ai/google-adk/generative-ui/a2ui/fixed-schema.md
19
16
  - https://docs.copilotkit.ai/reference/v2/hooks/useAgentContext.md
20
17
 
21
- Fetch all four pages in one step rather than one after another. Sequential fetching
22
- cost about nine minutes of the first proven journey.
23
-
24
18
  The documentation scope is `google-adk`. A `/adk/` URL answers with a redirect stub
25
19
  whose body is the canonical path, so fetch the `google-adk` form.
26
20
 
@@ -32,15 +26,9 @@ agent and not for ADK.
32
26
 
33
27
  Record the selected framework, vendor, model, required credential variable names, and
34
28
  these URLs.
35
- Do not use remembered CopilotKit instructions.
36
-
37
- A fetch tool that refuses a URL, or fails to reach it, reports a limit of the tool and
38
- not a fact about the page. Retrieve the same URL a second way before you judge it. Run
39
- `curl -fsSL <url>`, or read the same page without the `.md` suffix. Treat a page as
40
- unavailable only after a second method also fails.
41
29
 
42
30
  If the pages support the selection, run
43
- `npx --yes copilotkit@4.9.1 onboard read frontend/plan`.
31
+ `npx --yes copilotkit@4.9.4 onboard read frontend/plan`.
44
32
 
45
- If a second method also fails, or a page that loads does not support the selection, run
46
- `npx --yes copilotkit@4.9.1 onboard read unsupported/no-validated-path`.
33
+ If the documentation does not support the selection, run
34
+ `npx --yes copilotkit@4.9.4 onboard read unsupported/no-validated-path`.