$ loading_
Run autonomous code changes in a sandbox and return reviewable Git diffs.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "agentKimi" yet — see the docs or source repo.
Inspect the failing unit tests in the current git worktree, identify the issue, and fix the code. Make only the minimum necessary changes, then return a reviewable git diff and briefly explain the reason and impact.
A code diff that fixes the test failures, plus a brief explanation for human review before merging.
Scan this repository for duplicated user-permission validation logic, extract it into reusable functions or modules, preserve existing behavior, and return the git diff with refactoring notes.
A diff showing the refactor into reusable structures, with notes on compatibility and affected files.
Review the current service's input-handling code, look for possible command injection or path traversal risks, add necessary validation and safeguards, and return the git diff, risk summary, and verification suggestions.
A code diff that strengthens input validation and security safeguards, with identified risks and suggested verification steps.
Run local kimi-code CLI tasks remotely through QQ messages via MCP.
Execute code and search docs securely across sandboxed environments for development workflows.
Access Kimss agents, models, files, and vector stores from MCP clients.
Connect AI to Kimai for managing timesheets, projects, customers, and activities.
Search, validate, and cross-reference structured Markdown knowledge vaults for AI workflows.
Connect Korean search, maps, and routing APIs to LLM applications.