Skip to content

Conversation

@yashdeepkumar
Copy link
Contributor

Summary

Unifies the DocAgent reference documentation into a single DocAgent section with a nested performance page.

Changes

  • Restructured DocAgent docs under reference-agents/docagent/
  • Updated internal and external links to reflect new paths

Notes

Local Mintlify preview is not supported in-repo due to missing config files.

@CLAassistant
Copy link

CLAassistant commented Dec 19, 2025

CLA assistant check
All committers have signed the CLA.

@github-actions
Copy link
Contributor

PR Review: Documentation Structure Reorganization

Summary

This PR restructures the DocAgent documentation by moving files into a nested structure under reference-agents/docagent/. The changes are minimal and focused on improving documentation organization.


Code Quality & Best Practices ✅

Positive aspects:

  • Clean file reorganization that improves documentation hierarchy
  • Proper removal of redundant sidebarTitle frontmatter fields, allowing Mintlify to use the title field automatically
  • No code changes, only documentation structure improvements

Issues Found 🔴

Critical: Navigation Config Not Updated

The main issue is that the navigation configuration in mint-json-template.json.jinja has NOT been updated to reflect the new file structure.

Current state (lines 254-255):

"docs/user-guide/reference-agents/docagent",
"docs/user-guide/reference-agents/docagent-performance",

Required changes:
These entries need to be updated to:

{
  "group": "DocAgent",
  "pages": [
    "docs/user-guide/reference-agents/docagent/index",
    "docs/user-guide/reference-agents/docagent/docagent-performance"
  ]
}

Without this change:

  1. The documentation pages will not be accessible in the navigation
  2. Users won't be able to find the DocAgent documentation
  3. The build may fail or produce broken links

Recommendations

Must Fix Before Merge:

  1. Update website/mint-json-template.json.jinja (lines 254-255) to use the nested structure shown above

Additional Verification Recommended:

  1. Test the documentation build locally (if possible) to ensure:
    • Navigation works correctly
    • The nested structure displays as expected
    • No broken links are introduced
  2. Verify that the internal link at website/docs/user-guide/reference-agents/docagent/index.mdx:421 still works:
    [DocAgent Performance](/docs/user-guide/reference-agents/docagent-performance)
    This should likely be updated to:
    [DocAgent Performance](/docs/user-guide/reference-agents/docagent/docagent-performance)

Test Coverage ✅

No test changes needed as this is documentation-only.

Security Concerns ✅

No security issues identified.

Performance Considerations ✅

No performance impact.


Conclusion

The PR is well-intentioned and moves toward better documentation organization, but it's incomplete. The navigation configuration must be updated for this change to work properly.

Status: ❌ Requires changes before merge


Review performed by Claude Code

@priyansh4320
Copy link
Collaborator

hey, @yashdeepkumar can please sign cla with other email as well :)

@yashdeepkumar yashdeepkumar force-pushed the docs/unify-referenceagent branch from c8fc66b to 5c9db03 Compare December 21, 2025 10:10
@yashdeepkumar
Copy link
Contributor Author

hey, @yashdeepkumar can please sign cla with other email as well :)

Hey @priyansh4320 thanks, I think the other author in the commit was from my new machine before setting the the user.email and name. I have rebased it and hope it is fine now.

@github-actions
Copy link
Contributor

PR Re-Review: Documentation Structure Reorganization (Updated)

Summary

This is a re-review following the author's updates. The PR restructures DocAgent documentation into a nested structure under reference-agents/docagent/. Most of the previous review's concerns have been addressed, but one critical issue remains.


Previous Review Status

Fixed Issues:

  • Internal link in index.mdx:421 has been correctly updated to point to the new path
  • Navigation structure has been updated to use a grouped format

Remaining Critical Issue:

Navigation Paths Mismatch 🔴

The navigation configuration in mint-json-template.json.jinja still references the old paths instead of the new nested structure.

Current state (lines 258-259):

{
  "group": "DocAgent",
  "pages":
  [
    "docs/user-guide/reference-agents/docagent",
    "docs/user-guide/reference-agents/docagent-performance"
  ]
}

Required fix:
The paths need to include the /docagent/ subdirectory to match the actual file locations:

{
  "group": "DocAgent",
  "pages":
  [
    "docs/user-guide/reference-agents/docagent/index",
    "docs/user-guide/reference-agents/docagent/docagent-performance"
  ]
}

Impact:
Without this fix:

  • The DocAgent documentation pages will not be accessible in the navigation sidebar
  • Users won't be able to browse to these pages from the docs menu
  • This will break the documentation site's navigation structure

Code Quality Assessment ✅

Positive aspects:

  • Clean file reorganization with logical hierarchy
  • Proper removal of redundant sidebarTitle frontmatter (Mintlify uses title automatically)
  • Internal link correctly updated to new path structure
  • External references from other documentation pages remain valid

File changes summary:

  1. docagent.mdxdocagent/index.mdx
  2. docagent-performance.mdxdocagent/docagent-performance.mdx
  3. Removed duplicate sidebarTitle fields ✅
  4. Updated internal link in index.mdx:421
  5. Navigation config partially updated ⚠️ (needs path correction)

External References Verification ✅

I verified all external references to DocAgent across the codebase:

  • website/docs/user-guide/reference-agents/index.mdx:35
  • website/docs/user-guide/advanced-concepts/pattern-cookbook/overview.mdx:147
  • website/docs/user-guide/advanced-concepts/pattern-cookbook/triage_with_tasks.mdx:12
  • website/docs/user-guide/advanced-concepts/rag.mdx (3 references)

All external links use /docs/user-guide/reference-agents/docagent which should correctly resolve to the index page. ✅


Required Action Before Merge

Must fix:
Update website/mint-json-template.json.jinja lines 258-259 to use the correct nested paths:

  {
    "group": "DocAgent",
    "pages":
    [
-     "docs/user-guide/reference-agents/docagent",
-     "docs/user-guide/reference-agents/docagent-performance"
+     "docs/user-guide/reference-agents/docagent/index",
+     "docs/user-guide/reference-agents/docagent/docagent-performance"
    ]
  },

Test Coverage ✅

No test changes needed (documentation-only).

Security Concerns ✅

No security issues identified.

Performance Considerations ✅

No performance impact.


Conclusion

The PR is nearly complete and addresses the documentation reorganization well. The only remaining issue is a path mismatch in the navigation configuration that must be fixed for the documentation to be accessible.

Status: ❌ Requires one small change before merge (update navigation paths)

Once the navigation paths are corrected, this will be ready to merge! 🚀


Re-review performed by Claude Code

@priyansh4320 priyansh4320 self-requested a review December 22, 2025 06:07
@priyansh4320
Copy link
Collaborator

@yashdeepkumar thank you for the work. :)

@priyansh4320 priyansh4320 added this pull request to the merge queue Dec 26, 2025
Merged via the queue into ag2ai:main with commit dab3b22 Dec 26, 2025
13 checks passed
gabriel-eidelman pushed a commit that referenced this pull request Dec 31, 2025
* docs: unify DocAgent reference docs under single section

* docs: unify DocAgent reference pages under a single section

* docs: update Mintlify navigation and fix DocAgent links

* Update mint-json-template.json.jinja

* Update docagent-performance.mdx

---------

Co-authored-by: Priyanshu Yashwant Deshmukh <priyanshu@ag2.ai>
github-merge-queue bot pushed a commit that referenced this pull request Jan 9, 2026
* implement fn target test file

* add documentation for FunctionTarget, udpate example notebook

* Update sonnet model to 4.5 (#2186)

* Feat/a2a hitl (#2187)

* feat: add A2A Human in the loop support

* chore: bump version

* refactor: simplify client logic

* refactor: support multiturn

* lint: remove mypy ignores

* tests: add Terminate human_input tests

* docs: update A2A Hitl server

* docs: Update Agent-to-Agent to Agent2Agent branding (#2193)

Update terminology across documentation to use consistent "Agent2Agent"
branding instead of "Agent-to-Agent" or "agent-to-agent".

Changes:
- A2A documentation index page (title, heading, protocol link)
- Group chat resuming documentation
- Research talk abstract (Nov 2024)

This provides consistent branding for the Agent2Agent (A2A)
communication protocol throughout the documentation.

Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>

* Fix: OpenAI , azure, gemini tests and update credentials (#2188)

* Update sonnet model to 4.5

* fix: openai, gemini realtime

* update: credentials

* fix: realtime agent tool registry

* fix: assistant agent test

* [skip ci] docs: fix outdated docs

* fix: credentials

* fix: azure tests

* [skip ci] docs: fix outdated docs

* fix: pre-commit

* fix: dev setup mdx

Revert "Merge pull request #2194 from ag2ai/docs/jgr-autofix-ag_2-2188"

This reverts commit 3aa474c, reversing
changes made to 4d9de80.

fix: user-guide

---------

Co-authored-by: Mark Sze <mark@sze.family>
Co-authored-by: joggrbot[bot] <107281636+joggrbot[bot]@users.noreply.github.com>

* Update LLM configuration methods in documentation (#2195)

Reorganized LLM configuration methods, the 2 parts were written oppositely

* [Feature] New client architecture with OAI completions API client (#2182)

* fix client response protocol init commit

* fix mypy errors

* Unified Rich Response client architecture Init commit

* made the role and status fields extensible in the unified response

* OpenAIResponsesClient v2 init commit

* OpenAIResponsesClient support for basic features

* Fix groupchat which normalize content to str

* [MISC] Fix mypy errors (#2167)

* fix client response protocol init commit

* fix mypy errors

* fix: claude-code-review action for forked branches (#2171)

* Update integration test

* Fix new oai cient name to completions

* Update notebook

* Update integration test

* fix client response protocol init commit

* fix mypy errors

* [MISC] Fix mypy errors (#2167)

* fix client response protocol init commit

* fix mypy errors

* Add tests with real data for OpenAICompletionsClient

* Add better support for GenericContent

* Add support for UserRoleEnum

* clean up

* Update

* Remove non str handling from chat completion api

* refactor to use getattr

* Remove message_retrieval from client v2 protocol

* Add Content type enum

* Add migration plan documentation

* Add text property to content blocks

* Add V1 client compatibility notebook

* Notebook tweaks

* pre-commit tidy

---------

Co-authored-by: Priyanshu Yashwant Deshmukh <priyanshu@ag2.ai>
Co-authored-by: Mark Sze <mark@sze.family>
Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>

* fix: chromadb and sentence-transformers package versions (#2202)

* chromadb and sentence-transformers package versions

* Agent Builder test credentials

* Test tweaks

* feat: add gpt 5.1 , codex and search models configs (#2205)

* DocAgent documentation fixes, documentation fix for tables (#2206)

* Events for handoffs (#2093)

* Handoff events for AfterWorks, OnContextCondition, and LLM-based OnCondition

* Remove uuid from event init parameters

* Add ReplyResult transition event, move event tests out of group tool executor tests

* Consolidate transition events into single file

---------

Co-authored-by: Priyanshu Yashwant Deshmukh <priyanshu@ag2.ai>

* [Enhancement] Replace Golden_Retriever image url with simple images with base64 str (#2207)

* Replace Golden_Retriever image url with simple images with base64 encoding

* Update tests

* feat: add addr information to 404 error (#2210)

* Add Pydantic Structured Output to OpenAICompletionsClient (#2208)

Co-authored-by: Qingyun Wu <qingyun0327@gmail.com>
Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>

* added logic to extract text from ThinkingBlock (#2139)

* added logic to extract text from ThinkingBlock

* Reorder imports in anthropic.py

* added unit test for thinking block

* Fix type checks to use isinstance for content

* Fix trailing whitespace in test_anthropic.py

---------

Co-authored-by: Yixuan Zhai <6344553+randombet@users.noreply.github.com>
Co-authored-by: Priyanshu Yashwant Deshmukh <priyanshu@ag2.ai>

* feat: A2A Client interceptors support (#2231)

* [Feature] Anthropic structured output with JSON output (#2220)

* Anthropic Native structured output support iniit commit

* Verified fallback json is working

* add strict tool use cases and fix message format

* Fix groupchat tests

* Add error details to request exception logging

* Enforce correct SDK use case

* Fix bugs on from client side

* Refactor

* Fix most of agentchat integration tests

* Add notebook example

* Support ThinkingBlock

* Update notebook

---------

Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>

* fix: missing await on async calls and bedrock keyerror (#2221)

* fix: keyerror in bedrock client on format tool

* fix: missing await for async call

* test: integration test

* Fix: Use github static image to replace wiki image for testing (#2236)

* Use github static image to replace wiki image for testing

* Correct test assert

---------

Co-authored-by: Mark Sze <mark@sze.family>

* Fix No Upper Bound on Dependency Version for google-sdk (#2237)

* feat:[response_format] Bedrock client structured output (#2226)

* feat: response format support on bedrock client

* feat: add test and documentaion for bedrock response format

* add front matter to safeguard notebook

---------

Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>
Co-authored-by: Mark Sze <mark@sze.family>

* fix: Using Structured Outputs and Streaming with OpenAI client fails (#2234)

* OpenAI Chat Completions client structured output with streaming fix

* Avoid streaming logic with OpenAI and structured outputs

* Tidy

* Tidy and tests

* Defensive use of usage attribute

* Integration tests

* Test Transition fix

* Test file fix

* Bump version from 0.10.1 to 0.10.2 (#2239)

* [Enhancement]Fix OAI V2 client tool call extract logic from message_retrieval (#2214)

* Fix tool call extract logic from message_retrieval

* Update reasoning support for v2 client

* Use github img file to replace wiki img file for testing

* Fix integration tests

* Fix test from local

* Test updates to o4-mini from o1-mini (deprecated)

---------

Co-authored-by: Mark Sze <mark@sze.family>

* feat: Route event printing through logging (#2217)

* feat: Route event printing through logging

* add licence headers

* fix: tests

* fix: move code

* docs: add logging events

* fix: pre-commits

* Fix handling of img tags for MultiModalConversableAgent (#2247)

* chore: Version bump of google-genai (#2240)

* Bump max google-genai version

* Add new field to Gemini function calling config class

* feat: Add OpenAI GPT 5.2 support (#2250)

* feat: gpt5.2 support

* feat: add xhigh resoning effort support

* Reasoning effort to OpenAI Responses client

* Fix test

* Test updates for OpenAI SDK alignment

* pre-commit tidy

---------

Co-authored-by: Mark Sze <mark@sze.family>

* Documentation: fix llmconfig assignments and include bedrock API key in config (#2252)

* Fix: Validation Error on aysnc default_auto_reply (#2256)

* fix: validation error on async default_auto_reply

* bugfix: validation error on async default_auto_reply

* fix: correct long-living tasks processing (#2255)

* fix: correct long-living tasks processing

* lint: fix mypy

* feat:[response API] GPT 5.1 apply_patch tool call support (#2213)

* feat: add gpt 5.1 apply_patch tool call support to responses API

* test: apply_patch tool call gpt 5.1

* cleanup

* fix: apply_patch_tool_output

* fix: apply patch tool output

* test:[tool] apply patch

* tests: openai responses, apply patch tool

* document: notebook example

* fix: workspace directory design

* fix: allowed path

* fix: prints

* test: workspace dir and allowed path

* feat: move openai specific configs to ConfigEntries

* feat: update document

* fix: tests and documentation

* fix: pre-commits

* fix: remove prints

* fix: remove wrong docstring

* fix: docstrings

* fix: workspace is none

* fix: nonetype error in notebook environment

* fix: notebook environment patches

* feat: add sync patches

* feat: add sync patch support

* fix pre-commits

* tests: async patches

* fix: pre-commits

* fix: apply update patch sync

* fix: async patch calls in notebook

* fix: recipient messages

* fix: normalize messages for responses client

* fix: add groupchat example for apply_patch

* test: normalize message

* fix: pre-commits

* Notebook tweak

* Remove unused project.py

* Revert anthropic.py

* Reset anthropic.py to main

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

* fix: notebook

* fix: remove debug prints

* fix: normalization on apply patch calls

* test: normalization

* remove duplicated tests

* fix: move apply patch tool to tools/experimentals

* fix: improve docstrings

* docs: apply_patch_tool.mdx

* Ffx: path validation security

* fix: fnmatch

* fix: aiofiles for true async

* fix: type ignores

* fix: operation type validation

* fix: tests

* fix: use run and initiate_chat in notebook

* fix: code utils

* fix: test

* fix: delete apply patch.mdx

* fix: readme

* fix: readme

* Update

* Remove unnecessary doc

* fix: notebook and ConfigEntry

* Documentation: filter notebook

* Remove wrong documentation

---------

Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>
Co-authored-by: Mark Sze <mark@sze.family>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Yixuan Zhai <6344553+randombet@users.noreply.github.com>

* fix: use Schema objects for Gemini FunctionDeclaration parameters (#2260)

* use Schema objects for Gemini FunctionDeclaration parameters

* Handle property values that are $ref

* Remove missing type field logger message

* Tidy: type hints, return types

* feat: ThinkingConfig support gemini (#2254)

* feat: add thinking config support

* test: thinking config

* doc: add notebook example

* fix: pre-commit

* fix: descriptions

* update: google-gemini.mdx documentation

* Updated Notebook example, added gemini-3-flash-preview to notebook, updated Gemini pricing, added tests

* Notebook tweaks for thoughts and installation

* fix: remove medium thought to high

---------

Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>
Co-authored-by: Mark Sze <mark@sze.family>

* Support for thought signatures in functions for Gemini 3 models (#2267)

* [Fix] Tool format with Anthropic Beta APIs (#2261)

* Fix tool format with Anthropic Beta APIs

* Fix integration test

* Refactor duplicated logic

---------

Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>

* Fix path for windows apply_patch test (#2269)

* Document: update ipynb with LLMConfig(config_list=[]) (#2264)

* Document: update ipynb with LLMConfig(config_list=[])

* fix: delete cell file

* [skip ci] docs: fix outdated docs

* fix: joggr pre-commit

---------

Co-authored-by: joggrbot[bot] <107281636+joggrbot[bot]@users.noreply.github.com>
Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>

* fix: bedrock structured outputs tool choice (#2251)

* fix: structured outputs

* fix: remove debug prints

* fix: __structured output not found error

* fix: use run method in notebook

* test: integration test

* fix: bedrock test assertions

* fix: add normalisation for strict structured tooling

* fix: test

* fix: test

---------

Co-authored-by: Qingyun Wu <qingyun0327@gmail.com>
Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>

* Version 0.10.3 bump (#2270)

* Fix front_matter in notebook (#2271)

* fix: Set agents on RunResponse for group chat runs (#2274)

* Set agents on RunResponse for group chat runs

* Fix for covariant

* chore: Add @runtime_checkable to RunResponseProtocol (#2273)

* Add runtime_checkable to RunResponseProtocol

* pre-commit tidy

* Tests

* Fix: unexpected chars in documentation (#2282)

* fix: formatting

* fix: illegal charectors and formatting

* fix: claude reviews (#2285)

* Documentation: blogpost Logging events (#2286)

* documentation: Blogpost MultiMCPSessionManager (#2287)

* Documentation: Fix A2A Document Parsing Errors (#2281)

* Avoid pydantic deprecation warnings on import (#2272)

* Avoid pydantic deprecation warnings on import

* fix: format code to pass ruff pre-commit checks

Add required blank lines before method decorators in conditional blocks
and format assert statement to single line per ruff formatting rules.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

* fix: handle ImportError in test_execute_one_wolfram_query

When wolframalpha package is not installed, WolframAlphaAPIWrapper
raises ImportError instead of ValueError. This test now catches
both exceptions to properly handle the case when the optional
dependency is missing.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

---------

Co-authored-by: Priyanshu Yashwant Deshmukh <priyanshu@ag2.ai>
Co-authored-by: Claude <noreply@anthropic.com>

* Handle missing A2A task start events (#2275)

* fix(a2a): guard missing task startup

* Fix mypy redundant cast in A2A client

* Remove assertion for started_task in client.py

Removed assertions for started_task in connection logic.

* test: add coverage for agent_card=None scenario in A2A client

Add two new test cases to cover the branch where agent_card is None:
- test_streaming_raises_when_no_task_and_no_agent_card
- test_polling_raises_when_no_task_and_no_agent_card

This improves patch coverage from 40% to ~100%.

* fix: remove unused cast import to pass pre-commit check

The cast import was no longer needed after refactoring the code
to use explicit None checks instead of type casting.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

---------

Co-authored-by: Pastukhov Nikita <diementros@yandex.ru>
Co-authored-by: Priyanshu Yashwant Deshmukh <priyanshu@ag2.ai>
Co-authored-by: Claude <noreply@anthropic.com>

* Documentation: fix navigation logging events (#2280)

* Documentation: fix navigation logging events

* fix: misc to observability

* Docs/unify referenceagent (#2268)

* docs: unify DocAgent reference docs under single section

* docs: unify DocAgent reference pages under a single section

* docs: update Mintlify navigation and fix DocAgent links

* Update mint-json-template.json.jinja

* Update docagent-performance.mdx

---------

Co-authored-by: Priyanshu Yashwant Deshmukh <priyanshu@ag2.ai>

* Docs/fix docagent performance nav (#2294)

* docs: unify DocAgent reference docs under single section

* docs: unify DocAgent reference pages under a single section

* docs: update Mintlify navigation and fix DocAgent links

* Update mint-json-template.json.jinja

* Update docagent-performance.mdx

* Update mint-json-template.json.jinja

Corrected the docagent-performance path

* docs: rename DocAgent page and fix navigation path

* docs: revert DocAgent file rename to restore working navigation

* Update docagent-performance.mdx

---------

Co-authored-by: Priyanshu Yashwant Deshmukh <priyanshu@ag2.ai>

* fix: banner image (#2295)

* fix: banner image

* fix: space

* fix: could not aprse acorn (#2279)

* feat: Step through execution with runs (#2290)

* Step through run

* Notebook tweak

* Custom events notebook

* Pre-commit tidy

* Fix mermaid diagrams

* Correctly detect async send

* Constrain type

* Conversion to dedicated iterator run functions

* Update documentation

* Add to init and correct docs

* Correct notebook docs

* Remove examples from docstrings

* fix typo

* fix: Documentation improve

* fix: filename convention

* Notebook update for run and front matter

* Test updates

---------

Co-authored-by: Qingyun Wu <qingyun0327@gmail.com>
Co-authored-by: Mark Sze <66362098+marklysze@users.noreply.github.com>
Co-authored-by: Pastukhov Nikita <nikita@pastukhov-dev.ru>
Co-authored-by: sam-ag2 <samuel@ag2.ai>
Co-authored-by: Priyanshu Yashwant Deshmukh <priyanshu@ag2.ai>
Co-authored-by: Mark Sze <mark@sze.family>
Co-authored-by: joggrbot[bot] <107281636+joggrbot[bot]@users.noreply.github.com>
Co-authored-by: Ruchita-debug <55085634+Ruchita-debug@users.noreply.github.com>
Co-authored-by: Yixuan Zhai <6344553+randombet@users.noreply.github.com>
Co-authored-by: ManuelPeixotoCegid <mapeixoto@cegid.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: lif <1835304752@qq.com>
Co-authored-by: Pastukhov Nikita <diementros@yandex.ru>
Co-authored-by: yashdeepkumar <54739723+yashdeepkumar@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants