#ModGitHub stars

1. 97publish-pulsechenyuxiaojin/cyxj-notchStatus line showing days since your last publish (traffic light) and the next scheduled release, read from Markdown files under NOTCH_PUBLISH_ROOT; also writes ~/.claude/notch/pulse.json for the cyxj-notch panel.3
2. 98qrxqrxcodes/qrx-mcpBranded, print-ready art QR codes that always scan. Describe a picture and a link; QRX paints the code, verifies it decodes, and hosts it on a qrx.to short link.3
3. 99specialists-uixtrm-dev/specialistsTeammate-style transcript rows for Specialists in Claude Code: channel wakes, the fallback wake, and Specialists MCP tool calls under native labels. Companion to the specialists plugin, which raises those rows and therefore cannot draw them itself.3
4. 100superlcmyu381792/superlcmOne local conversation archive for Claude Code, Codex, Pi and dsh harness. Keep full originals, write background layered summaries, recall exact sources and continue tasks across tools. Connections default to native host compaction. Claude Code takeover is an explicit optional setting. Open /superlcm:console to manage connections and summary models.3
5. 101where-were-wescarrillo/agent-pluginsTimestamps your prompts in the transcript and adds up/down prompt navigation to the footer3
6. 102dependency-lensYeonwooSung/my-claude-code-modsDependency Lens1.1k
7. 103deploy-verifyescapeboy/claude-code-kitAfter a deploy command, runs 5 post-deploy checks from the project's .claude/deploy-verify.json and tells Claude the result.97
8. 104agent-compact-advisorapolenkov/agent-compact-advisorStatus line score 0-100 for how good a moment it is to /compact now, with reasons; offers a ready /compact that keeps goal, decisions, leftovers and paths, and adds that template to every compaction. Never compacts by itself.2
9. 105angry-switchertimoncool/angry-switcherSwear at Claude Code all you like: every prompt is cleaned before Claude reads it. Typos, translit, the wrong keyboard layout, swearing and CAPS go; the meaning, paths and numbers stay; your original rides along so nothing is lost.2
10. 106bash-trimcpeaustriajc/cj-stackCuts a long successful Bash output to its first 2,000 and last 3,000 characters with a count of the cut lines; read-like commands, failures and saved outputs pass whole2
11. 107clef-model-routerAbelNavarro/clef-model-routerPicks the Claude model and reasoning effort for each turn with Cloudflare Clef, a fast decision model on Workers AI.2
12. 108cs-radioben-rogerson/claude-counter-strikeCounter-Strike 1.6 radio calls for Claude Code: 'Fire in the hole' on deploys, 'Bomb has been defused' when a long turn lands. /radio to toggle.2
13. 109folioHT0710/claudzillaDraws the reply's frame and divider lines at the chat's live width, so they never wrap, and links file paths to the margin2
14. 110grudgehacksurvivor/claude-modsEvery look you reject becomes a rule for that project, with your screenshot, so Claude never brings it back; it also says when Claude changed UI without a preview you approved2
15. 111inline-imageschrisns/claude-image-cli-modSee the images that commands print with the iTerm2 inline images protocol (imgcat and friends), the images Claude sends you, and the Mermaid diagrams Claude draws (Wardley maps too), in the Claude Code transcript. Real pixels in iTerm2, kitty and Ghostty; coloured blocks everywhere else.2
16. 112jeffortvntrungld/jeffortPer-turn effort routing with TypeSafe's Jev, applied per request at turn.step. Off until /jev on.2
17. 113lossless-compactMusicStudioNYC/lossless-compactLossless compaction for Claude Code: nothing is summarized; everything removed is archived verbatim with provenance and restorable by id (Jev or a local ruleset). /lossless to inspect and restore.2
18. 114multiharness-delegationfranciscocarloserra/claude-code-multiharness-delegationRoute Agent calls with subagent_type multiharness-delegation:<harness> to an interactive external harness in tmux2
19. 115open-pathskiankhooban/open-pathsClick a file path in a Claude Code reply and the file opens in its app: Word, Preview, Excel, VS Code for code. Not a Finder reveal.2
20. 116path-linkspepperonas/path-linksClickable paths in Claude Code replies: click a folder to open it in your file manager, click the file name to open the file2
21. 117personasSApplefeld/claude-kitPIANO-esque cognitive layer: persona-keyed state, heartbeat liveness, controller with classify/actuate2
22. 118quarryjackboykin/quarryAdds Exa's results to WebSearch, and saves each WebFetch page to disk with the lines that answer the prompt2
23. 119shell-shimMichaelP17/claude-modsFixes the zsh pitfalls Claude's bash-style commands trip over — = words, unmatched globs, word splitting, aliases that hide programs, a missing timeout — before the command runs2
24. 120skill-pickiuliandita/modsNames the 1 to 5 installed skills that fit each prompt, as many as the ask needs, with a small model and, when TYPESAFE_API_KEY is set, a Jev score per skill; the log shows how often the main model then loads one. Default mode dry only measures. With a key set, the prompt and skill descriptions go to api.typesafe.ai (Jev) from every directory unless jevPaths limits it.2
25. 121smartcompactambervdberg/smartcompactA Claude Code mod that compacts the conversation at a good moment, asked for by the session after a nudge or chosen by a small judge model, and then tells Claude to continue.2
26. 122team-orchestratorherman925/925-cc-pluginsGuided team setup and a native workspace for rosters, verified session connections, task assignment and outcome review. Existing classic Orca teams remain available separately.2
27. 123tool-visibility-controllerrezzminator/tool-visibility-controllerClaude Code plugin for per-agent tool visibility — hide and refuse subagents, skills, MCP and built-in tools per loop2
28. 124translate-for-claudeJansenChang/claude-terminal-translateTranslate Chinese prompts to English before they reach Claude, and Claude's answers back to Chinese, with DeepSeek2
29. 125command-shelfYeonwooSung/my-claude-code-modsCommand Shelf1.1k
30. 126delay-no-moreruihe774/cc-delay-no-moreChoose which deferred tools (built-in and MCP) are loaded up front instead of waiting behind ToolSearch
31. 127promptcolorjwhall/promptcolor/promptcolor changes the background color of your own prompt lines in the Claude Code transcript so they are easy to spot while scrolling back.
32. 128session-relaunchlperezmo/session-relaunch/relaunch restarts this Claude Code session in a new terminal tab with --resume <id>, so MCP servers added since it started are loaded.
33. 129blast-radiusYeonwooSung/my-claude-code-modsBlast Radius1.1k
34. 130activity-logdiegorv/claude-mods-diegorvEvery event of the session, tool calls and their full results included, appended to a JSONL file per session1
35. 131agents-skillsyuu1111/claude-agents-skillsLoad Agent Skills from the shared .agents/skills directories (project and user) into Claude Code1
36. 132annotantneolitec/code-buddyComment on your dev app in the browser; one background agent per comment makes the change and answers in the thread1
37. 133architecture-treescholarly360/claude-mod-architecture-treeWhen a prompt uses architecture or design words, asks Claude to answer with the system architecture as a box-drawing tree.1
38. 134auto-effortjustmytwospence/claude-auto-effortSets the effort for each prompt: Jev rates how much reasoning the request needs, within the bounds (free by default, or capped at the session's own effort).1
39. 135bitranoxbitranox/bitranox-skillsBitranox skill collection: software-engineering workflow skills (planning, debugging, code review, clean architecture, language and tool references, humanizing) plus a self-improvement loop that captures session learnings into memory and CLAUDE.md, and PostToolUse hooks that validate JSON/YAML/XML edits and flag AI-writing tells in prose, feeding issues back for correction.1
40. 136bookmarksDazzleML/claude-bookmarksBookmarks and vim-style marks inside Claude Code conversations: mark a line, jump back, keep a reading position, browse your prompts, and let Claude cite earlier places as clickable links.1
41. 137break-reminderangomedia/claude-modsA countdown to your next break under the message box, and a gentle reminder after 50 minutes of work. Time away counts as a break1
42. 138calm-enhancekulovema2012/calm-modeCalm Enhance: an Enhance button above the prompt that turns a short, vague draft into a precise prompt from the recent conversation ("do 1 and 2" becomes the actual options). Press again to revert. Adapted from cc-prompt-enhance-mod by Yigit Budak (MIT).1
43. 139ciel-moderuihe774/cc-ciel-modeciel-mode (Call-in-CEL mode) for Claude Code: the model writes a small sandboxed CEL program that calls MCP and read-only tools, filters and joins their results, and returns only what it needs, so large tool outputs never fill the context. Every call still goes through the permission checks.1
44. 140classify-hintcorrupt952/dotfilesClassifies the user's prompt with Haiku against a few yes/no questions in one call and attaches a rough, non-binding hint for the main agent1
45. 141cleaner-tool-linesmacwilk/cc-extensionsDraws each tool call as one line, so the conversation reads straight through. Click a line to open the call.1
46. 142clear-resumem4cd4r4/clear-resumeWrite a handover when a Claude Code session gets long, type /clear, and the next session picks up where you left off with nothing to paste.1
47. 143collapse-workFrogAi/XenopusFolds a finished turn's work under a "Worked for" row, leaving the final answer, like Codex1
48. 144comfyui-localanderson-spider/claude-modsGenerate images, video and audio, keep workflows in your library, and manage node packs and restart the server on your self-hosted ComfyUI (and ComfyUI-Manager) through its HTTP API, using the url plugin option; unofficial, with slash commands ported from Comfy-Org/comfy-skills.1