1. 241forward-compactionrezzminator/sub-agent-compactThe model writes its own frontier through one tool and the plugin installs it at Claude Code's compaction, with facts recorded from typed events and every later message kept verbatim.
  2. 242godmodetinkybeatz/tinky-claude-modsSwitches a session into an agentic development loop: a PM main thread that plans tickets and dispatches dev and review agents
  3. 243habloJarvizx/habloHear Claude Code's replies aloud in your language: a Listen button under each reply and /hablo. Free and offline on macOS and Windows.
  4. 244harness-nachoNachofranco5/claude-harnessReglas de trabajo de Nacho + tablero al arrancar + skills (tarea, cerrar-tarea, planificar, proyecto-nuevo, cierre-charla, podar, plugins-proyecto, automejora) y subagentes (explorador, verificador) + guardianes (cierre, lectura cara, versión atrasada)
  5. 245history-keeperJvrd97/claude-history-keeperKeep a copy of every Claude Code session that the 30-day cleanup cannot delete, as JSONL and readable markdown, with message times in the transcript and /history to list and search old sessions.
  6. 246idea-agentsSpendry/claude-code-modsSix exploration agents per project (premise-breaker, field-crosser, idea-sizer, and a lazy trogg for each), one ideas/ folder per run, and a guide in Claude's system prompt for running a round
  7. 247idle-opacityUnlimitedxIQ/brysons-claude-setupWezTerm window goes more solid when Claude stops, so the answer is easier to read, and back to see-through while it works
  8. 248image-previewezoushen/claude-modsRender image paths inline: a thumbnail in the user's prompt row (aspect-fill, max 10 rows) and in agent reply rows when the model mentions a path. Recognises a file path ending in a supported image extension (quoted or with escaped spaces too), or an [Image #N] token for a pasted picture.
  9. 249image-shrinkervarunmoka7/image-shrinkerShrinks big screenshots before Claude reads them, so long sessions last longer, and tells Claude the scale so clicks still land in the right place.
  10. 250jc-gearboxKevinYu1580/jcGearboxPicks the model once per session and the effort every turn, using TypeSafe's Jev to judge how hard the task is.
  11. 251jev-judgekuhku/kuhaku-pluginsAnnotates Claude Code's AskUserQuestion dialog with TypeSafe Jev's probabilities and a recommended option (labels in Japanese).
  12. 252jev-model-routerm-esm/jev-model-routerTypeSafe Jev scores each prompt and switches the session's model and effort to match, per turn.
  13. 253jev-model-routerTolsee/dotfilesSets the model of every subagent spawn (agent.spawn). The agent type sets the baseline: agents whose definition pins a model keep it, general-purpose, claude and Plan get sonnet, and TypeSafe's Jev may move those down to haiku, or up to opus when the task is risky (production, money, irreversible data). Sonnet is the ceiling otherwise; a description starting with [opus] goes to opus. No key or no verdict in time leaves the baseline, never the parent's Opus.
  14. 254jev-routerXvectorio/xvector-claude-modsPer-turn model routing: TypeSafe Jev picks Haiku, Sonnet or Opus for each prompt (port of gargpratyush/jev-router).
  15. 255jev-vault-gatesohryuu101/jev-vault-gateUses Jev to spot durable knowledge in a turn and auto-capture it into ~/vault/raw/ for /compile.
  16. 256jev-watchdeadczarvc-labs/jev-watchLogs what fast-jev-compaction does (outcomes, TypeSafe requests, compactions, hook failures) to JSONL files.
  17. 257kittexbrenocq/kittexShows the LaTeX math in Claude's replies as real equations: images in kitty and Ghostty, text elsewhere
  18. 258lanesjysvai/lanesOn-demand lane sessions beside your main Claude Code session, from the first idea to release: a mid lane for side work that needs judgement (research, plans, reviews, tests), a low lane for search, scripts and crawling. Works on the Claude API, Amazon Bedrock, Vertex AI and Foundry. Status line, /lane, and tools the main model calls.
  19. 259latexRafeSymonds/claude-modsReads your LaTeX header's macros so math renders in the desktop app, and draws math as Unicode in the terminal
  20. 260latex-inlinedazebug/latex-inlineShows the LaTeX math in Claude's replies typeset: as pictures inline with the text in Ghostty, cmux and kitty, and as Unicode text in other terminals
  21. 261laymanvarunmoka7/laymanSay "I'm lost" and Claude explains its last reply again in everyday words. A Claude Code plugin.
  22. 262legion-mod-runnerdnh33/legion-modCompanion to Legion Mod: starts, resumes and stops Legion's agents for it. Install both; it does nothing on its own.
  23. 263lemonclaudeadamczhang/LemonClaudeSwitch a Claude Code session between Claude and a local Lemonade model mid-conversation: /lemonade lists, downloads and switches models
  24. 264living-skyanillaksu/living-skyA living sky behind Windows Terminal that follows Claude: stars at night, clouds by day, a meteor shower while it works, a storm when a tool fails, a sunrise when it is done.
  25. 265lofi-modelakmadev/claude-modsLo-fi beats while Claude works: a soft loop starts on long turns and stops the moment it's done. /lofi toggles.
  26. 266m3l-prompt-optimizermonte3l/m3l-prompt-optimizerA Claude Code plugin that refines the prompts you type using Anthropic's prompt-engineering guidance, and passes clear prompts through unchanged
  27. 267mc-pvp-botsFatihBastan/mc-pvp-botsFight Minecraft PvP bots while Claude works, and get handed back the moment it's done or needs you.
  28. 268md-promptmtmtian/cc-modsPaints Markdown onto Claude Code's prompt box as you type: fenced code becomes a syntax-highlighted card (even before you close the fence), and inline code, bold, italic, lists, task lists, quotes, tables, links and headings are styled. Paint only, your text is never changed.
  29. 269mermaidWhiteKr/claude-plugin-mermaidRender ```mermaid fences in assistant replies: Unicode box-drawing diagrams in the terminal (lovely-mermaid), light/dark vector SVG on desktop, VS Code and mobile (beautiful-mermaid + dagre)
  30. 270mermaid-inlinedazebug/mermaid-inlineDraws the Mermaid diagrams in Claude's replies: as pictures in Ghostty, cmux and kitty, and as Unicode box drawing in other terminals
  31. 271message-timestampsderekwden-droid/message-timestampsShows the time above each prompt you send and each reply Claude starts, in the terminal and the desktop app's Code tab, as clock times or as time since. /timestamps switches between clock, relative and off.
  32. 272mod-doctorTihi321/claude-mods/mod-doctor checks every installed mod in one report: hook scripts, plugin mods loaded, Node, Claude Code version, errors and ESLint timings in the mods log.
  33. 273model-advisorBlockyogi/model-advisorJudges each prompt with Haiku 5.5 and, when a different Claude model or effort would help, offers Best, Balanced and Keep before the prompt runs.
  34. 274model-routerAkurganow/model-routerRoutes each subagent to a Claude model and effort that fits its task, judged by Jev. For each routed Agent call it sends the task text and the agent's description, cut to 500 characters and with secrets masked, to api.typesafe.ai, openrouter.ai or both, as its provider option allows, each with its own key. The scope option picks the calls: core routes the four built-in agents, extended also routes user, project and plugin agents and sends their text. A config tool opens the plugin's own dialog to change both options. A key set as an environment variable is visible to every process Claude Code starts.
  35. 275model-routerclaudemodz/modsRuns chosen subagent types on cheaper models (Explore on Haiku by default) and adds /router to switch the main session between fast, smart and unchanged. Shows the active model beside the spinner.
  36. 276model-routeriddhi-sulakshana/claude-modsClassifies each turn as mechanical, ordinary or hard and routes it to a matching model and reasoning effort, logging every decision.
  37. 277model-routermuhx/model-routerJev (TypeSafe) picks the model and effort per task type; ported from jev-model-router with main-loop model switching on
  38. 278model-routerredjackfred/claude-code-modsRuns Explore subagents on haiku and general-purpose ones on sonnet unless a model was asked for
  39. 279model-routershezis/claude-model-routerRoutes each turn and subagent to a cheaper model when the task is simple, never above the model you picked
  40. 280model-routervladdu/claude-kitPicks Haiku, Sonnet or Opus from a conversation's first prompt and keeps it until /clear; /router on|off
  41. 281model-routerwesleystagg/model-routerPicks the model and effort for every subagent by having your session's own model judge the task, instead of a fixed model per agent type. /router to pause, check, or read its log.
  42. 282model-router-slimGambetaClub/model-routerPicks the model and the reasoning effort for each task with TypeSafe's Jev, a System One decision model reached either through TypeSafe's own API or the Vercel AI Gateway. Routes subagent models at agent.spawn and the main loop's effort and model at turn.step, in both directions, falling back to the engine's built-in classifier when no key is set.
  43. 283modwallRadTech-Solutions/modwallA firewall for other mods: shows what each mod can reach as it loads, and audits, holds or refuses it by your policy
  44. 284moodcodewe-are-digital-architects/moodcodeClaude Code, now with feelings: an evolving mood, shown as a face beside short, simple, human replies whenever it changes.
  45. 285msgate-agentsJustAzul/agent-cli-sdkmsgate-agents: ready-made agent types (second opinion, code review, cross-check, expert persona, delegation, ad hoc) that the msgate engine plugin answers with Codex, plus the dispatch skill.
  46. 286nerd-fontAH-Merii/claude-pluginsWorks out once whether a Nerd Font is installed, with a setting to force it on or off, so mods that draw Nerd Font glyphs know when to fall back
  47. 287new-projectDanielGarcia-Carrillo/claude-improvements/new-project bootstraps a repository from the claude-improvements baselines, in Claude Code or Codex: proposes the kickoff answers from your description and asks only the open ones, copies the web, Rust or Python template, writes CLAUDE.md and the stack ADR, then installs and runs verify
  48. 288nuetmmedum/nuetA dim time above each prompt and each reply in the Claude Code transcript, written the way your own clock setting writes it