Update dependencies and enhance security and functionality

[geffzhang]

This pull request introduces several improvements and updates across documentation, CI configuration, and the .NET codebase. The main highlights are the addition of DeepSeek as a first-class native LLM provider (with full documentation and setup instructions), the introduction of the AgentQi Mobile companion app in docs and site maps, and important codebase improvements for thread safety and resource management. It also removes legacy Microsoft Agent Framework migration paths and clarifies configuration guidance.

Documentation updates:

• DeepSeek is now documented as a native LLM provider, with instructions for setup, environment variables, and model selection in README.md, docs/QUICKSTART.md, and docs/GETTING_STARTED.md. [1] [2] [3] [4] [5] [6]
• The AgentQi Mobile companion app is now referenced in the main README, docs map, and site map, including a direct link and description as an open-source Android operator console. [1] [2] [3] [4]

Microsoft Agent Framework (MAF) migration cleanup:

• Removes legacy/experimental MAF migration documentation and clarifies that only the supported configuration section should be used, both in English and Chinese documentation. [1] [2] [3] [4] [5] [6]

Codebase improvements:

• Renames MafExecutionContext to AgentExecutionContext and moves it to a more appropriate namespace, adding a TryGetCurrent() method for safer context access. [1] [2]
• The agent system prompt builder now includes optional IDENTITY.md and MEMORY.md workspace files for richer agent context.
• The process execution backend now ensures that spawned processes are killed on cancellation, preventing orphaned processes and resource leaks.
• OpenClawToolExecutor is made thread-safe by locking around shared state, and tool declaration access is now atomic and safe for concurrent operations. [1] [2]

CI/CD updates:

• Updates the GitHub Actions setup-node action to the latest commit hash for improved security and reliability. [1] [2]

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 482df9cc-687b-4f10-bf13-d308226fe1fd

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch main

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}