[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-75847":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":25,"hasPages":25,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":29,"readmeContent":30,"aiSummary":31,"trendingCount":16,"starSnapshotCount":16,"syncStatus":32,"lastSyncTime":33,"discoverSource":34},75847,"clawpatch","openclaw\u002Fclawpatch","openclaw","Review code. Patch bugs. Land PRs.","https:\u002F\u002Fclawpatch.ai",null,"TypeScript",735,107,3,1,0,4,34,541,20,10.1,"MIT License",false,"main",true,[27,28],"bot","review","2026-06-12 02:03:36","# 🩹 clawpatch\n\nAutomated code review that lands fixes.\n\n`clawpatch` maps a repo into semantic feature slices, reviews each slice with a\nprovider, persists findings, and can run an explicit fix loop for one finding at\na time.\n\nCurrent status: early CLI. Review\u002Freport\u002Fstate are implemented; patching exists\nbehind `clawpatch fix --finding \u003Cid>` and still requires manual review of the\nresulting worktree changes.\n\n## Install\n\n```bash\npnpm add -g clawpatch\n```\n\nFrom source:\n\n```bash\npnpm install\npnpm build\npnpm link --global\n```\n\n## Workflow\n\n```bash\nclawpatch init\nclawpatch map\nclawpatch review --limit 3 --jobs 3\nclawpatch review --mode deslopify --limit 3\nclawpatch ci --since origin\u002Fmain --output clawpatch-report.md\nclawpatch report\nclawpatch next\nclawpatch show --finding \u003Cid>\nclawpatch triage --finding \u003Cid> --status false-positive --note \"covered by tests\"\nclawpatch fix --finding \u003Cid>\nclawpatch open-pr --patch \u003CpatchAttemptId> --draft\nclawpatch revalidate --finding \u003Cid>\nclawpatch revalidate --all --status open\n```\n\n`fix` does not commit, push, open PRs, or land changes. It runs configured\nvalidation commands and records a patch attempt under `.clawpatch\u002F`.\n\n## What It Maps Today\n\n- npm package bins\n- selected root and workspace package scripts: `start`, `build`, `test`,\n  `lint`, `typecheck`, `format`\n- Node\u002FTypeScript workspace packages under `apps\u002F*`, `packages\u002F*`, and package\n  workspace patterns\n- package-less Node\u002FTypeScript app roots under monorepo folders such as\n  `apps\u002F*` and `packages\u002F*` when source or positive framework signals are\n  present\n- generic extension\u002Fplugin packages under workspace roots such as `extensions\u002F*`\n  and `plugins\u002F*`, including package metadata, source, docs, and nearby tests\n- semantic Node source groups for large packages, including runtime, commands,\n  auth, storage, monitor, webhook, setup, server, and client slices\n- Nx project metadata from `project.json`, including project-scoped validation\n  targets\n- Turborepo task metadata for workspace-aware validation commands and feature\n  context\n- Next.js `app\u002F` and `pages\u002F` routes, including routes inside monorepo apps\n- React Router routes and React components\n- Go package slices from `go list .\u002F...`, including command packages\n- Go package tests and same-repo imports as review context\n- Java\u002FKotlin Gradle source groups and root Gradle build\u002Ftest commands\n- JVM semantic roles from Java and Kotlin code evidence such as annotations,\n  imports, interfaces, inheritance, supertypes, and method signatures\n- Kotlin Android semantic roles for UI entrypoints, ViewModels, data\n  boundaries, external clients, and dependency injection, including Metro\n- C#\u002F.NET projects from `.sln`, `.slnx`, `.csproj`, `.fsproj`, and `.vbproj`\n  files, with conservative `dotnet build` \u002F `dotnet test` defaults\n- ASP.NET Core controllers, minimal API endpoints, C#\u002FF#\u002FVisual Basic source\n  groups, and .NET test projects\n- Ruby project metadata, executables, source groups, RSpec\u002FMinitest suites\n- Elixir Mix\u002FPhoenix projects, contexts, Phoenix web slices, runtime config,\n  Ecto migrations, project scripts, and ExUnit suites\n- Rust `src\u002Fmain.rs`, `src\u002Fbin\u002F*.rs`, `src\u002Flib.rs`, `crates\u002F*`, and\n  `tests\u002F*.rs`\n- C\u002FC++ standalone `main()` files, CMake `add_executable` \u002F `add_library`\n  targets, and autotools `bin_PROGRAMS` \u002F `lib_LTLIBRARIES` targets\n- Python project metadata, console scripts, bounded source groups, pytest suites,\n  and Flask\u002FFastAPI\u002FDjango routes\n- SwiftPM `Sources\u002F*` targets and `Tests\u002F*` suites\n- Laravel\u002FPHP projects from `composer.json` and `artisan`, including routes,\n  controllers, form requests, Artisan commands, jobs, services, models,\n  migrations, seeders, Composer scripts, and PHP test suites\n- common project config files\n\nDeeper framework mappers and agent-assisted enrichment are next steps.\n\n## Provider\n\nThe default provider is the local Codex CLI.\n\n```bash\ncodex --version\nclawpatch doctor\n```\n\nProvider calls use `codex exec` with strict JSON schemas. Review and revalidate\nrun read-only; fix planning runs with workspace-write because Codex may edit the\nworking tree during the explicit fix command.\n\nSet `CLAWPATCH_CODEX_SANDBOX` to override the Codex sandbox passed by\nClawpatch. Use any Codex sandbox mode, or `bypass`\u002F`none` to pass\n`--dangerously-bypass-approvals-and-sandbox` when the host environment already\nprovides isolation.\n\nSupported provider names today:\n\n- `codex`: local Codex CLI\n- `acpx`: any ACP-compatible coding agent (Codex \u002F Claude \u002F Pi \u002F Gemini \u002F ...) via openclaw\u002Facpx\n- `grok`: local Grok Build CLI\n- `opencode`: local OpenCode CLI\n- `mock`: deterministic test provider\n- `mock-fail`: failure test provider\n\n## Commands\n\n- `clawpatch init`: create `.clawpatch\u002F`, detect project basics, write config\n- `clawpatch map`: write feature records\n- `clawpatch status`: show project, dirty state, feature\u002Ffinding counts\n- `clawpatch review`: review pending or selected features\n- `clawpatch review --mode deslopify`: review only for locally provable slop cleanup\n- `clawpatch ci`: initialize if needed, map, review, write a report, and append a GitHub step summary\n- `clawpatch report`: print or write a Markdown findings report\n- `clawpatch next`: print the next actionable finding\n- `clawpatch show --finding \u003Cid>`: inspect one finding with evidence and suggested validation\n- `clawpatch triage --finding \u003Cid> --status \u003Cstatus>`: mark a finding with optional history note\n- `clawpatch fix --finding \u003Cid>`: run the explicit patch loop for one finding\n- `clawpatch open-pr --patch \u003Cid>`: commit an applied patch attempt and open a GitHub PR\n- `clawpatch revalidate --finding \u003Cid>`: re-check one finding\n- `clawpatch revalidate --all`: re-check open findings with report-style filters\n- `clawpatch doctor`: check provider availability\n- `clawpatch clean-locks`: clear feature locks\n\nUseful flags:\n\n- `--root \u003Cpath>`\n- `--state-dir \u003Cpath>`\n- `--config \u003Cpath>`\n- `--json`\n- `--plain`\n- `--limit \u003Cn>`\n- `--jobs \u003Cn>`\n- `--source \u003Cheuristic|auto|agent>`\n- `--feature \u003Cid>`\n- `--project \u003Cname-or-root>`\n- `--finding \u003Cid>`\n- `--status \u003Cstatus>`\n- `--severity \u003Cseverity>`\n- `--provider \u003Cname>`\n- `--model \u003Cname>`\n- `--reasoning-effort \u003Cnone|minimal|low|medium|high|xhigh>`\n- `--skip-git-repo-check`\n- `--output \u003Cpath>` \u002F `-o \u003Cpath>`\n- `--dry-run`\n- `--force`\n\nUnknown flags fail fast.\n\n## State\n\nState is project-local by default:\n\n```text\n.clawpatch\u002F\n  config.json\n  project.json\n  features\u002F*.json\n  findings\u002F*.json\n  patches\u002F*.json\n  reports\u002F*.md\n  runs\u002F*.json\n```\n\nFeature records are the durable work units. Findings and patch attempts link back\nto features so runs can resume and be audited.\n\n## Safety\n\n- Review does not edit files.\n- Fix is explicit and selected by finding ID.\n- Fix refuses a dirty source worktree by default.\n- Clawpatch commits, pushes, and opens PRs only from explicit patch commands such as `open-pr`.\n- Clawpatch does not land changes today.\n- Provider output is parsed through strict schemas.\n- Symlinked directories and generated build output are skipped during mapping.\n\nSee `docs\u002Fspec.md` for the longer product and implementation spec.\n","clawpatch 是一个用于代码审查、修复错误并提交拉取请求的自动化工具。它使用 TypeScript 构建，能够将代码仓库映射为语义化的功能切片，并通过提供商对每个切片进行审查，记录发现的问题，并支持针对特定问题运行修复循环。项目目前处于早期CLI阶段，已经实现了审查、报告和状态管理功能，但修复功能仍需手动审查结果。适用于需要提高代码质量和维护效率的开发团队，特别是那些希望利用自动化手段减少人工审查负担的场景。",2,"2026-06-11 03:53:30","CREATED_QUERY"]