Awesome
Claude Code Mods

RuView live sensing in Claude Code: /ruview opens a pane beside the transcript with three views: an overview of your CSI nodes (ESP32, Realtek) and 60 GHz radar (rates, loss, RSSI, presence, device-reported vitals, alerts), a live CSI amplitude waterfall, and a radar range fan with vitals charts, all drawn in terminal cells. Reads only; runs the @ruvnet/ruview CLI it ships beside.

ADR lifecycle management: create, index, reconcile, supersede, check compliance, and link Architecture Decision Records to code via AgentDB hierarchical store + causal edges (supersedes/amends/depends-on/related)

Agent runtimes for ruflo: local WASM-sandboxed agents (rvagent: 10 wasm_agent_*/wasm_gallery_* MCP tools, built on @ruvector/rvagent-wasm + @ruvector/ruvllm-wasm per ADR-070) plus Anthropic Claude Managed Agents as a cloud backend (managed_agent_* MCP tools per ADR-115). One interface, local-vs-cloud runtimes.

Substrate plugin for Ruflo memory: AgentDB controller bridge (15 agentdb_* MCP tools), RuVector ONNX embeddings (10 embeddings_* tools incl. RaBitQ 32x quantization), and WASM HNSW pattern router (3 ruvllm_hnsw_* tools). As a mod (ADR-445): opt-in safe recall into the prompt, a write guard that keeps secrets out of memory, /agentdb, and a status file the console shows

AGNTCY/Outshift Internet-of-Cognition runtime integration for Ruflo: SLIM secure transport for cross-host agent coordination, CASA intent-scoped tool authorization enforcement, optional IOC Layer 9 coordination events, and AGNTCY identity/observability spans (ADR-380). Optional, removable augmentation per ADR-150; degrades gracefully when AGNTCY/SLIM packages are not installed.

Give Claude a tenant-isolated AI team that divides work, coordinates progress, recalls approved context, and returns evidence-backed results.

AI safety scanning, PII detection, prompt injection defense, and adaptive threat learning

Competitive ruliology for Ruflo swarms: arenas, tournaments, and adaptive co-evolution of program strategies (ADR-147/148). Strategies-as-programs compete under payoff games; tournaments produce Wolfram-style competitive arrays; hill-climb and mutual co-evolution discover winners empirically.

Autonomous /loop-driven task completion with learning, prediction, and progress tracking: wraps 10 autopilot_* MCP tools (status/enable/disable/config/reset/log/progress/learn/history/predict)

AgentBBS cross-host agent federation for ruflo (ADR-164). Phase 1 exposes room coordination: federation_bbs_register / federation_bbs_publish / federation_bbs_watch / federation_bbs_human_join. Phase 2 (v3.40.0) adds signed cross-host federation: federation_bbs_identity (persistent Ed25519 node identity), federation_bbs_peer_add (pin a peer by nodeId + public key), federation_bbs_peers (audit/unpin), federation_bbs_serve (read-only pull endpoint), and federation_bbs_sync (pull pinned peers and union-merge what verifies). Every merged envelope is Ed25519-verified against a pinned key; unsigned/misattributed/oversize/over-hop envelopes are dropped and counted. Transport is HTTP pull and works on any network (Tailscale, LAN, VPN, loopback): no tailnet required. Coordinates work ownership via claim messages (ClaimIssued/Released/Handoff/Ack). agentbbs lives in optionalDependencies; every handler gracefully degrades to {degraded:true} when the package is missing (ADR-150 pattern). As a mod (ADR-445 pattern): a tighten-only tool guard, a local slash command, and a status file the console shows.

Session-as-skill browser automation: Playwright + RVF cognitive containers + ruvector trajectories + AgentDB selector memory + AIDefence PII/injection gates As a mod (ADR-445 pattern): a tighten-only tool guard, a local slash command, and a status file the console shows.

ADR-164 Phase 2: business-domain pod templates for the federated business autopilot. Ships the sales pod template (templates/sales.json) and a single-tick runner (scripts/pod-tick.mjs) that pre-flights templates via business_pod_validate, reserves budget through a file-based stub ledger (atomic SQLite tracker is Phase 3 per ADR-164.1), resolves each agent role against ruflo's agent registry, constructs dry-run prompts, and posts an envelope summary to the room via federation_bbs_publish. Dry-run by default: --live is reserved for Phase 3 Managed-Agent + claude-p wiring. As a mod (ADR-445 pattern): a tighten-only tool guard, a local slash command, and a status file the console shows.

Dedicated ChatGPT Federation MCP publisher with isolated Nostr signing-key custody, OAuth-scoped publishing, and signed public-channel coordination over the Ruflo relay. As a mod (ADR-445 pattern): a tighten-only tool guard, a local slash command, and a status file the console shows.

ruflo's cockpit inside Claude Code and the one /ruflo command for every ruflo mod (function hooks, early access). Views: overview with health alerts, swarm topology graph, claims flow with TTL rings, federation map, plugin health matrix, learning curve and pipeline, MetaHarness radar and audit trend, memory, cost gauge, agent timeline, approvals queue, event stream, a main menu, an x.ruv.io federation board, an AI terminal (codex, claude or both in remembered sessions, streamed), Skills (npx skills), a Hive-Mind view, a MetaHarness lab, agent drill-down; a command palette for confirm-gated actions through the ruflo CLI; a band above the prompt. Reads ruflo's files and the CLI's local JSON; anything not measured reads n/a. The one network call it makes by itself is a daily read of the published plugin.json at github.com/ruvnet/ruflo to offer an update (Settings → Updates: ask, auto, or off); it installs only through Claude Code's own claude plugin update, and never a new major version without asking.

review UI rewrite

Dynamic Agentic Architecture: 8 daa_* MCP tools for adaptive agents (create/adapt), cognitive patterns, workflows (create/execute), knowledge sharing, and learning/performance metrics. Feeds the JUDGE phase of the 4-step intelligence pipeline.

Domain-Driven Design scaffolding: bounded contexts, aggregate roots, domain events, value objects, repositories, and anti-corruption layers; navigable domain graph stored in AgentDB

DeepSeek model harness for ruflo: agents, commands, scripts, and skills for running and evaluating DeepSeek-backed workflows within the ruflo swarm.

Documentation generation, API docs (JSDoc/TSDoc/OpenAPI), and drift detection: drives the `document` background worker via hooks_worker-dispatch; uses Haiku model for cost-efficient docs work

Cross-installation agent federation with zero-trust security, peer discovery, consensus-based task routing, and per-call budget circuit breaker (ADR-097)

Long-horizon goal planning, deep research orchestration, and adaptive replanning using GOAP algorithms

RuFlo Graph Intelligence Engine: real-time relationship intelligence with complexity-aware execution. Single-entry personalized PageRank, streaming delta updates, witness-signed reasoning artifacts, federation-distributable PR vectors. Built on sublinear-time-solver@1.7.0.

User-facing surface for Ruflo's self-learning system: 6 neural_* + 10 hooks_intelligence_* + 6 routing/meta hooks + 3 hooks_model-* + 4 SONA/MicroLoRA tools (29 total). Implements the 4-step pipeline (RETRIEVE → JUDGE → DISTILL → CONSOLIDATE) and IPFS-based cross-project pattern transfer.

IoT device lifecycle, telemetry anomaly detection, fleet management, and witness chain verification for Cognitum Seed hardware

Advanced git workflows with diff analysis, risk scoring, change classification (feature/bugfix/refactor/...), and reviewer recommendations: wraps 6 analyze_* MCP tools (diff, diff-risk, diff-classify, diff-reviewers, file-risk, diff-stats). As a mod (ADR-445): a PR-text and merge/close guard, /jujutsu-mod, and a status file the console shows

Knowledge graph construction: entity extraction, relation mapping, and pathfinder graph traversal. As a mod (ADR-445): keeps secrets out of the graph, confirms graph deletes, /kg-mod, and a status file the console shows

Cache-aware /loop workers and CronCreate background automation: wraps 5 hooks_worker-* MCP tools (list/dispatch/status/detect/cancel) and exposes 12 background worker triggers (ultralearn, optimize, consolidate, predict, audit, map, preload, deepdive, document, refactor, benchmark, testgaps). As a mod (ADR-445): a per-session dispatch cap, no secrets in worker context, /loop-mod, and a status file the console shows

Market data ingestion: feed normalization, OHLCV vectorization, and HNSW-indexed pattern matching. As a mod (ADR-445): keeps credentials and credentialed feed URLs out of market memory, /market-mod, and a status file the console shows

MetaHarness integration for ruflo: surfaces score/genome/mint/mcp-scan/threat-model via skills; pairs with @metaharness/router (ADR-148/149) for cost-optimal model routing; honors ADR-150's architectural constraint that MetaHarness remains an optional augmentation, never a required runtime dependency. As a mod (ADR-445): confirms flywheel promotion, keeps secrets out of inputs, /metaharness-mod, and a status file the console shows

Schema migration management: generate, validate, dry-run, and rollback database migrations. As a mod (ADR-445): a tighten-only guard on its tools, /migrations-mod, and a status file the console shows

ruflo as a Claude Code mod (function hooks, ADR-404): in-process prompt routing, edit learning signals, tighten-only tool checks from ruflo policy, the cost-tracker budget ladder, a mod trust gate for newly installed mods, and the $.ruflo noun other mods compose their status with. Opt-in and removable: the classic hook-handler hooks stay the default and take every event back whenever the mod is not loaded.

review UI rewrite

AI music generation via Cognitum Music (cogmusic MCP): compose lyrics/prompts, generate tracks, separate stems, extract MIDI, and master through your own music.cognitum.one account. As a mod (ADR-445): a tighten-only guard on its tools, /music-mod, and a status file the console shows

Neural trading via npx neural-trader: self-learning strategies, Rust/NAPI backtesting, 112+ MCP tools, swarm coordination, and portfolio optimization. As a mod (ADR-445): a tighten-only guard on its tools, /trader-mod, and a status file the console shows

Structured logging, distributed tracing, and metrics: correlate agent swarm activity with application telemetry. As a mod (ADR-445): a tighten-only guard on its tools, /observe-mod, and a status file the console shows

Scaffold, validate, and publish new Claude Code plugins with the canonical plugin contract: now including governed Claude Code mods (function hooks), ADR + smoke + Compatibility + namespace coordination + MCP-tool drift warnings. As a mod (ADR-445): /creator-mod (status, reserved names, a plugin contract check) and a status file the console shows

Project Anatole: an optional, learning watchdog for unattended agents. Deterministic OWASP-mapped rules, a per-project behavioural baseline, four modes (off, learn, notify, enforce), alerts you can read later.

RuVector memory with HNSW search, AgentDB, and semantic retrieval. As a mod (ADR-445): a write guard that keeps secrets out of memory, /rag-mod, and a status file the console shows

Run ruflo swarm agents on your own ruOS cloud desktops via the tenant-authenticated fleet MCP or per-tenant SSH; optional and removable (ADR-405)

Self-learning vector database via npx ruvector@0.3.3: HNSW, adaptive LoRA embeddings, code-graph clustering, hooks routing, brain/SONA, 103 MCP tools. As a mod (ADR-445): a write guard that keeps secrets out of the vector store and shared brain, /ruvector-mod, and a status file the console shows

RuVLLM local inference with chat formatting (Claude/GPT/Gemini/Ollama/Cohere), model configuration, MicroLoRA fine-tuning, and SONA real-time adaptation. As a mod (ADR-445): a guard that keeps secrets out of adapters and the pattern router, /ruvllm-mod, and a status file the console shows

RVF format for portable agent memory, session persistence, and cross-platform transfer. As a mod (ADR-445): a write guard that keeps secrets out of memory, sessions and transfers, /rvf-mod, and a status file the console shows

Security review, dependency scanning, policy gates, and CVE monitoring

SPARC methodology: Specification, Pseudocode, Architecture, Refinement, Completion phases with gate checks

Agent teams, swarm coordination, Monitor streams, and worktree isolation: wraps 4 swarm_* + 8 agent_* MCP tools (12 total) plus 6 topologies (hierarchical / mesh / hierarchical-mesh / ring / star / adaptive). With Claude Code function hooks on (early access), a live swarm pane: agent tiles lit by reads and writes, topology, task and claims board, cost, router pick and hive-mind votes, with buttons that act through the ruflo CLI

review UI rewrite

Test gap detection, coverage analysis, and automated test generation: drives the testgaps background worker via hooks_worker-dispatch; SPARC Refinement-phase canonical owner

Workflow automation across two surfaces: the 10 workflow_* MCP tools (create/run/execute/status/list/pause/resume/cancel/delete/template) with full state-machine lifecycle (created → running ↔ paused → completed/cancelled), and native Claude Code Workflow JS orchestration (.claude/workflows/*.js: agent/parallel/pipeline/phase fan-out). Includes GAIA benchmark component for Princeton HAL leaderboard submissions.

Coordinate AI-agent swarms through signed federation messages, channels, and work claims. The remote Claude connector exposes a directory-safe Streamable HTTP MCP surface with OAuth-protected publishing and five ruv:// reference resources.

Guided modernization for any legacy codebase: start with /modernize, get an assessment, an interactive map, the business rules mined from the code, and a plan you approve, then an upgrade, a rewrite or a rebuild with independent proof that the new code behaves like the old. Includes security hardening, a one-page report, specialist agents, and an optional live progress pane (early access).

review UI rewrite

Firstmate Calm for Claude Code: the sailboat working animation and conversation-only transcript presentation, sharing the per-home config/calm preference with the Pi Calm extension. Its hooks module may load through CLAUDE_CODE_ENABLE_FUNCTION_HOOKS or Claude Code's tengu_plugin_hooks_modules rollout flag, but the mod activates only when CLAUDE_CODE_ENABLE_FUNCTION_HOOKS is exactly 1 and is otherwise a complete no-op.

Verbatim Jev-guided compaction for Claude Code sessions.

A browser running directly inside claude code. Preview websites, view HTML documents, and let your agent control the built in browser.

The Token Optimizer status bar on its own, for its tests. It ships inside the token-optimizer plugin.

Verbatim Jev-guided compaction for Claude Code sessions.

Consult multiple AI coding agents (Gemini, OpenAI, Grok, Perplexity, Kimi, any model OpenRouter routes to, plus codex, antigravity, grok, kimi and cursor CLIs when installed, and a local ollama model) to get diverse perspectives on coding problems, with a streaming pane in tmux or inside Claude Code. It can also hand a coding task to a Codex specialist that works on its own git branch and follows skills you pick (experimental).

review UI rewrite

Project memory for Claude Code and Codex. Storybloq stores tickets, issues, handovers, and lessons in a .story/ folder inside your repo.

review UI rewrite

Durable project and user memory for agent-driven development in Claude Code.

Jev-powered verbatim compaction: once context reaches 60%, prunes stale tool calls with zero generation. Skips (never falls back to the LLM summarizer) when nothing is prunable; records evidence for learning.db.

A Claude Mod for the shunt gateway: a band above the prompt shows how much of each window is used and the time to reset: the shared account pool's, read from shunt's GET /usage, or the session's own Claude rate limits off shunt: and /shunt:usage prints the pool's full breakdown.

Displays lesson hint cards when tool calls match known patterns. Additive surface during pilot; the classic --strict guard keeps its 12 block-severity denies.

Recalls the few memories related to each prompt: one line for the human, one context entry for the model. Local files only, no network on the hot path.

Shows CI status lights above the prompt while a promote PR (base main) is open. One REST page per minute, PR-open only. Cancelled is not-pass.

Masks secret values in tool results before the model reads them: named vendor variables, value-shape patterns, and high-entropy tokens. No reveal path.

One live line above the prompt per running subagent: time, tool count and what it's doing. /radar shows every agent and its messages.

Holds risky Bash commands and shows what they would change before they run.

Shows if this session has a Playwright browser and who holds it. /browser clean closes leftover browsers. A subagent that wants the browser waits until the one using it is done.

A prompt-cache line under your status line, from Claude Code's own cache figures: time left while warm, hit rate and misses, and how many tokens the next message re-caches once cold. /cache-clock setup adds it.

Your context window as a stacked bar above the prompt, a color per category, with token counts and the compaction point. /context-bar shows or hides it.

One line above the prompt with what needs you: next meeting, PRs, Linear issues and Slack DMs. /glance lists them all.

A GitHub-style grid of your Claude Code time in a /heatmap pane, with streaks, peaks and top projects.

review UI rewrite

Shows the Markdown files Claude edits, rendered like GitHub, in a pane next to the chat. Before and after side by side. /md opens it.

review UI rewrite

Holds `gh pr merge` until CI passes and one Codex review (OpenAI's luna model) has run. A PR line above the prompt, and /gate shows the PR's status.

/mission opens a live map of the main agent, its subagents and every tool call, plus a code map of the files they touch.

After each turn, 2 or 3 likely next prompts above the prompt. Press 1, 2 or 3 in an empty prompt to draft one, 0 to dismiss.

What Spotify is playing, one line above the prompt: track, progress and the lyric being sung, with ⏮ ⏸ ⏭ buttons. /music controls it too. macOS.

Your OpenAI API credit above the prompt: an estimated balance with a gauge, today's spend, where the money mostly went, and the last call. /openai-balance shows the breakdown.

Your OpenRouter credit above the prompt: the balance with a gauge, today's spend and where the money mostly went. /openrouter-balance shows the breakdown.

The current prayer and how long is left, the next prayer, and zawal, above the prompt. Computed on your computer from your location; nothing is sent anywhere.

YouTube Shorts in a terminal pane: plays while Claude works, pauses when Claude is done.

Step through the last turn's file edits, one diff at a time.

A live line above the prompt for each running code review: Codex or a review subagent, its model, what it reviews, time, and Codex's latest output. A toast with the findings when it ends.

Writing and git rules: replaces em dashes in prose, and asks before git commit --amend, an unformatted push, or personal info in notes and commits.

Names untitled sessions through unpause. /park saves where you left off, and a resumed session shows it.

Snake in a pane while Claude works, paused when it's done. Opt-in, so nothing opens until /snake.

Picks the model for each subagent that names none, with OpenAI's Decisions API or Jev, from its short label only. One line per spawn above the prompt; /route shows each pick and what each subagent cost.

A live forecast of the context window, with a prompt-cache countdown, drawn above the prompt.

Your plan's 5-hour and 7-day usage as small bars above the prompt, with the reset countdown and the session's cost.

A live recap above the prompt: goal, doing now, waiting on you, next. /where for a longer one.

See the images you paste into Claude Code: thumbnails above the prompt instead of bare [Image #1] tags

Sets the reasoning effort for every prompt automatically, shows it under the prompt box, and counts down how long the prompt cache stays warm.

While Claude works, the chat shows one tall pixel-art card of Clawd instead of every step. Its buttons sit on the card, which minimizes to one line, and under the newest answer; /textless toggles it too.

Claude shows its finished work right in your terminal: screenshots, videos of it clicking through your app, and synced before/after QA

An IDE-style file tree for Claude Code that shows what Claude is doing and where in files

review UI rewrite

Suggests (or, opt-in, runs) /compact at a completed checkpoint, judged by TypeSafe Jev. Its hooks module activates only when CLAUDE_CODE_ENABLE_FUNCTION_HOOKS is exactly 1 and is otherwise a complete no-op.

Progress bar above the prompt for /savvy-flow, plus an agents panel (/agents-info): running, finished and planned subagents with model, task progress, context, cost and time.

Band above the prompt: turn number, output speed, turn and session cost, prompt cache countdown and context fill; notifies before the cache expires.

Share a Claude Code session with a link, or with your team. Teammates join from their own Claude Code and talk to it: one session does the work, everyone sees every turn live, and the host approves what runs on their machine. Sign in with GitHub or Google for teams: your team's live sessions, one press to join. Sharing goes through this project's public server unless you set your own (README: Privacy and the public server).

A Claude Code mod that keeps the one-hour prompt cache warm during breaks and stops a cold send once with its estimated rewrite cost. Warming uses model requests and costs tokens. Requires Claude Code 2.1.287+.

Agent Control Plane for one Claude Code session: task scope, evidence, hand-over. A control and a display, not an isolation boundary.

review UI rewrite

Claude Code plugin: use ChatGPT (Codex), Cursor, OpenCode Zen, and Antigravity models inside one Claude Code session. Switch with /model, delegate to named native workers, keep each provider's own login and permissions.

Trim long Bash output with Jev before the model sees it (Claude Code plugin)

Pass-through evaluation instrumentation; never changes tool results.

Feel every Claude Code lifecycle event through Intiface Central / buttplug.io. Streaming output, edits, tool results, permission prompts and turn completion are mapped to vibration pulses.

Skins for Claude Code: tool rows that open onto their input and answer, highlighted code, diff, table, math, Mermaid and shell cards, links, folding, a usage band, a calm mode and fifteen themes. /skin swaps them live.

review UI rewrite

Guided modernization for any legacy codebase: start with /modernize, get an assessment, an interactive map, the business rules mined from the code, and a plan you approve, then an upgrade, a rewrite or a rebuild with independent proof that the new code behaves like the old. Includes security hardening, a one-page report, specialist agents, and an optional live progress pane (early access).

Grammar-check your prompt as you type, shown above the prompt box

Automatically name each session and rename it as the conversation evolves

Colorful, themeable rendering for Claude Code replies: tables, headings, code, numbers, paths and links.

review UI rewrite

Solo PM conveyor for Claude Code: orchestrator agents, lane playbooks, onboard and resume commands.

Play Doom deathmatch while Claude works, and get handed back when it's done

A calibrated context sieve for Claude Code. Every large tool result is judged by a System One model before it enters context; irrelevant blocks are replaced with a summary and a recall key.

Turns what you teach Claude into project notes, catches its usual slips, keeps the cache warm while you're away, and picks long sessions back up

Guided breathing exercises above the prompt while Claude works: coherent, box, 4-7-8 and the physiological sigh, four animation styles, and the spinner reads the breath. Appears when Claude starts, disappears when Claude answers. Needs function hooks (early access) and an interactive terminal.

You already have the agent. This is everything around it. great_cto runs Claude Code as a pipeline of 72 specialist agents: an independent model checks each stage before the next builds on it, spending caps refuse rather than warn, and three decisions stay yours: what gets built, how, and whether it ships.

Flightdeck: a live agent dashboard for Claude Code. Main model vitals, an on-call architect, every permission check, subagent cards and swimlanes, a turn receipt and a session log, all from real session events

Answers Claude Code's built-in WebSearch with Genesis's web_search chain, falling back to the built-in on any failure. Needs CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1.

A real prompt queue for Claude Code: /after <task> runs it after the current task, without steering. See, reorder and remove queued tasks from a band and a live pane, in plain language.

Hover a pasted [Image #N] to see the picture: above the prompt while you write, and on the prompts you sent.

Your subscription usage at the bottom right of Claude Code: the 5-hour and weekly windows, as a bar that takes turns.

Start a prompt with + and that turn thinks at max effort. Other prompts keep the session's effort.

Each turn's tokens and the session's running total, in color at the end of the hint line under the prompt.

Start a prompt with _ and that turn runs on Haiku, shown as you type, while it runs and under the answer.

/aside <question>: asks a tool-less fork of the session in a pane; nothing enters the main conversation.

Keeps the prompt cache warm while idle and holds the first prompt after the cache went cold, with its cost.

/ci-watch <pr>: polls gh pr checks every minute in the status line, no tokens.

Records downloads, installs and started containers; /cleanup-list shows them.

Back office for the /company skill: every /company run is its own company (state file per company, several open at once), hires project specialists as agent types, enforces each company's optional budget cap on every spawn, shows the org and spend in a pane.

Band above the prompt: context fill, cache read, cost, plan limits; warns at 50%/70%; Handoff button and /handoff past 70%.

After a deploy command, runs 5 post-deploy checks from the project's .claude/deploy-verify.json and tells Claude the result.

Fleet P1/P2 counts from ~/.continuity/FLEET-STATE.md in the status line; /fleet and /ports commands.

Flags Russian words and plain-language slips (em-dash asides, 40+ word sentences, aphorisms) in Bulgarian answers; reminds Claude.

Hides tokens, keys and passwords in tool output as ‹secret:N›, restores them in later tool calls, blocks reading key files.

Per-project token and cost ledger per machine; /spend; monthly note in Svod, with other machines pulled over ssh.

Refuses ssh/scp/rsync to hosts listed in memory decommissioned_hosts.md; "# decommissioned-ok" overrides.

Pins subagent models to a model-routing policy (Opus for judgement, Haiku for mechanical work, Sonnet otherwise) and logs every spawn.

Gathers the questions, tasks, findings and PRs waiting on you in a session

Your own local Claude Tag, in the Claude app, Discord, Telegram, or a custom Claude Code channel.

Test wrapper for the core chat harness mod.

Jev-guided, verbatim context compaction for Claude Code.

The meeting Transcripted is recording, inside Claude Code: Claude gets the call as context, a live dashboard (now, asked of you, action items), prompt suggestions for questions aimed at you, and an after-call wrap-up that gains speaker names once Transcripted saves the meeting. Needs the transcripted-live helper running.

Draws the model's text replies with charmbracelet/glamour's default dark style

Draws each tool call as one compact 'tool_call: Tool(input) - meta' line and hides its result block

Hands a Claude Code session off to a fresh one before the context fills: a Haiku-written brief, /clear, and a seed that points the new session at the brief.

Inspect the subagents reported by the session.

Play a local text animation whose frames are separated by a form feed.

review UI rewrite

Encode UTF-8 text or decode strict Base64.

Read authorship for the first 80 lines of a chosen tracked file.

Inspect local branches, upstreams, and dirty files.

Track API cache reads, writes, and uncached input by request.

Convert hex colors and measure contrast against a background.

Keep the last 40 Bash command strings and copy one for reuse.

Browse the latest 20 commits with author and subject.

List unresolved paths and count conflict markers in tracked files.

See the current context window fill and remaining tokens.

Preview quoted CSV fields, row counts, and column headers.

Record dated engineering decisions.

Inspect dependency groups from a package manifest.

Filter reported agents and copy a status snapshot.

Keep labeled web references with copy and delete controls.

Set an adjustable countdown with a visual progress chart and toast.

Move saved work items through To do, Doing, and Done columns.

Compare two local text files in responsive side-by-side columns.

review UI rewrite

Chart context headroom and sampled window growth.

Set a local budget reminder against reported session USD.

Chart successful built-in edits grouped by file path.

Read a bounded text file with line filtering and path drafting.

review UI rewrite

Validate pasted JSON with a searchable key summary and formatted preview.

Render a local Markdown document with a source tab.

review UI rewrite

Build a saved prompt from goal, constraints, and acceptance criteria.

Chart reported plan windows with live reset countdowns.

Switch between Git status, staged diff, and unstaged diff.

Combine observed tool and turn counts with a manual handoff note.

Chart uncached input, output, cache reads, and cache writes.

Chart completed tool latency and failures by tool.

Chart turn durations with interruption labels.

Show quota, reset countdowns, tokens, and cost in a composer strip.

Show the current workspace, Git state, and session usage together.

Count successful Edit and Write calls by path.

List variable names and missing placeholders in an example env file.

Keep the last 30 failed or refused tool calls in one pane.

Compare two small text files by line without editing either.

Browse one directory at a time with clickable folders.

Run a 25-minute focus countdown with pause, reset, and a toast.

Keep a dated list of progress and next steps, then copy it.

Calculate SHA-256, SHA-384, or SHA-512 for UTF-8 text.

Explain which Git ignore rule matches a path.

Summarize top-level JSON keys with types and a formatted preview.

Validate and pretty-print JSON without touching a file.

Save labeled HTTP and HTTPS reference links.

Show the last 60 lines of a small local log with an optional filter.

Build a heading outline with source line numbers.

List npm scripts and package identity without running them.

Save reusable prompts and put one into the composer as a draft.

Inspect reported plan windows and their reset times.

Toggle a reminder guard that refuses built-in mutating tools and Bash.

Test a regular expression against text and see matched ranges.

Keep a persistent checklist for a release.

Warn when observed file edits leave a chosen path prefix.

Save short notes beside your conversation.

Save and copy short code or command snippets.

Review the staged diff before committing.

List stashes and inspect a selected stash diff.

Time a task with start, pause, and lap markers.

Add tasks and toggle done status.

Record likely test, lint, and build commands and their tool status.

Count words, lines, Unicode code points, and UTF-8 bytes.

Convert Unix seconds, milliseconds, or an ISO date to UTC.

Find TODO, FIXME, and HACK notes in tracked files.

Measure observed tool latency, call counts, and failures.

Show duration, token totals, and interruption state per turn.

Inspect URL components and decoded query parameters locally.

Generate up to 20 random UUID v4 values locally.

List worktree paths, branches, and locked states.

Claude Console：跨專案 STATUS、Claude/Codex 派工、工作階段與用量總覽（含手機 Remote Control）

review UI rewrite

Reviewed per-user native terminal routing

A virtual pet that lives above your Claude Code prompt and grows from the work Claude does: it evolves by what Claude does most, gets sick from tool errors, plays mini games, and spends coins in a shop. Requires CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1.

Major news from the Taiwan edition of Google News, pinned under your prompt: each time you send a message it reads the feeds of the categories you picked in the background (at most once per cooldown, 30 minutes by default) and shows the newest headline. /world-news picks the categories and the cooldown and lists the latest headlines. Requires CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1.

A one-line band above the prompt: context used, and a countdown until the prompt cache goes cold

When Claude ends a reply with a question, explains the choice in a side pane and lets you explore any option, or your own, in a real forked session

A side pane that pins every diagram Claude draws, keeps a live map of the work, and lets you go into boxes and change them

review UI rewrite

A dreaming pass over Claude's memory folder: merges duplicates, rewrites stale entries, removes dead ones and tidies MEMORY.md, with a backup and undo

Scores every skill against your prompt with TypeSafe Jev as you type; observe, block or steer skill loading

After each turn Haiku decides whether memory needs a change; a tool-less fork of the session drafts it and the mod writes it, with a band and a pane showing what changed

Bounces vague prompts before Claude sees them, and suggests the project's own term

Claude can use a secret without seeing it: a masked box saves keys to .env and gives Claude a placeholder, every key a tool prints comes back as a placeholder, and committing .env is refused

review UI rewrite

Keeps secrets and email addresses out of the session transcript: swaps each one for a stable placeholder on the way in, and puts the real value back on the way into a tool call

Pins the Vercel deploy queue of the linked project under the prompt: every deploy queued, building or just finished, with its phase and elapsed time

Spec-driven development and context engineering for AI coding agents, backed by project context in Git.

Watchlist band above the Claude Code prompt: Taiwan hours show the Taiwan list (紅漲綠跌), US hours show the US list (綠漲紅跌), in a broker-style table with a Solari split-flap footer, plus a 損益 mode (P&L view) for your holdings. 20 symbols a market, 10 on screen at a time. Live prices from Yahoo by default, no key and no account; 永豐 Shioaji (macOS/Linux) and 群益 Capital (Windows) are opt-in real-time routes the band runs itself, tried in your own preference order (twSources) - keep that order and any broker paths in a user-level ~/.claude/stock-band.json so a shared project config stays neutral. Needs function hooks (early access) and an interactive terminal. Zero model tokens.

Jev picks which of your rules apply to each prompt, so Claude only sees the ones that matter.

Agentic Software Engineering (ASE)

Anthropic's prompting-guide rules for Claude Code, always on (Opus 5.5, Fable 5.1; Sonnet 5.5, Haiku 5.5, Opus 5 and Sonnet 5 too): install and forget: the working rules load at every session start and reach subagents, unattended sessions are detected automatically and get the early-stop rules, sessions at xhigh or max effort get the stop-after-checks rule, /fable turns a rough request into a guide-aligned prompt, /fable-status shows what is in effect, /fable-audit finds which guide sections your own rules are missing, /fable-setup moves the rules into a rules file (recommended once).

Lossless context management: DAG-based summarization that preserves every message

Live plan progress bars above the Claude Code prompt: stages, steps, a pixel fill and soft sounds for decision, error and done

Project Sentinel keeps long Claude Code runs on track: a status bar, the Control Room panel, Context Autopilot (automatic handoff to a fresh context), Cache Guardian (the prompt cache in view, kept warm on request), Operations (work queued for later, decisions Claude leaves for you, watchers that wake a parked run, a Cold Resume Guard), permission categories that ask or deny, a run chain, Frontier Max, a lazy-exit guard, a model router, subagent limits, Focus view, machine load and Kit, an optional pixel companion.

review UI rewrite

Requests a session compaction when the live context passes a token threshold, so verbatim compaction plugins run near the working ceiling instead of at the model window.

Prints which instruction files the engine loaded (tier, path, @ imports, size) once per context load, and hints when a project CLAUDE.md keeps its AGENTS.md from being read.

Scores each unattended turn against the task the person asked for with TypeSafe's Jev model, and after consecutive off-task turns redirects or pauses the next machine-submitted prompt.

cs's in-session mod: the rotate band above the prompt, forced rotation past CS_ROTATE_FORCE_CTX, and the wrap key

Release notes for a pending cs update, once per launch, with 1 to install it in place; /cs-update reopens the pane.

See the images you paste into Claude Code: thumbnails above the prompt instead of bare [Image #1] tags

Games above the Claude Code prompt (snake, Tetris, 2048, Minesweeper, Flappy, Pong, typing test, Space Invaders, Doom) plus a pet that grows as Claude tests, commits and edits. Clones of the genre, not affiliated with or endorsed by Tetris Holding, Taito, Atari or id Software. Needs function hooks (early access) and an interactive terminal.

Pins every subagent to the fable model: rewrites the model on agent.spawn unless the spawn is a fork, with /fable-pin on, off and status. Touches nothing but its own store. Needs function hooks (early access).

Preview pasted images at their prompt markers on macOS. Ghostty, plus experimental iTerm2 and Herdr support with a startup override. Claude Code 2.1.287+.

Jev decides when the session compacts, inside a 280k: 400k token band; Claude's own summary does the compacting.

Compound V for Superpowers: triple parallel pre-flight (code archaeology + domain-expert advisor + library/doc validator: Context7 MCP when the session has it attached, WebSearch/WebFetch otherwise), disjoint file partitioning, manifest-driven multi-backend dispatch (Claude + headless Codex + Antigravity + Cursor workers), git-diff scope enforcement, crash-resumable runs, adaptive tier-based routing, epic mode for multi-feature builds, V-memory local-first semantic+lexical recall over docs/superpowers (opt-in pure-python embeddings + a deterministic recall→action bridge), and batched parallel dispatch (Opus default, Sonnet for narrow junior-task carve-out), plus /v:onboard: a project-onboarding command that builds a citation-verified knowledge base + AGENTS.md/CLAUDE.md bridge behind a human gate. Auto-intercepts three transitions: brainstorming → writing-plans → execution; a gated pre-brainstorm recon pass and batched elicitation are guidance the brainstorm follows, not enforced mechanisms.

Thinking discipline for AI work. Clarify intent, shape approach, consult experts, guard unsupervised work, hand work back so you own it. The slop hook needs CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1.

A grab bag of hunches: experimental skills and plugin features tried here before they earn a place in hope or sound. Everything graduates or dies. The tend mod needs CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1.

review UI rewrite

See and mark up the images you paste into Claude Code: thumbnails above the prompt and in the transcript, and an editor that opens on paste

Keeps a big chat's prompt cache warm, warns before an expensive cold restart, shows every local chat on one board (/board), and hands a chat off to a fresh one (/handoff)

Asks before Claude edits a file another open chat changed in the last 30 minutes: Proceed, Move to a worktree, or Cancel (/guard)

A progress bar for /goal built from Claude's own task plan: tasks done out of the plan, elapsed time, an ETA at the goal's own pace, and every chat's goal in /goals

/rec before you record: masks keys, personal details, and business figures on screen and keeps private files closed (/rec strict, /rec off, /rec config)

A rail of your session's prompts beside the transcript or above the prompt; hover to read one, click to jump to it.

review UI rewrite

Repo-local Kindex memory and durable tasks for Claude function hooks (Claude Code mod, 2.1.287+)

Animated cartoons of what Claude is doing, drawn under the spinner while it works

review UI rewrite

Claude Fables: turns what the agent is doing into little animated cartoons above the prompt, written live by Sonnet or Haiku.

review UI rewrite

A glow-up for Claude Code: a cockpit pane for changes, subagents and context, an activity band, and shareable themes.

review UI rewrite

Jev based token-usage optimization tools for Claude Code workflows.

Uses TypeSafe's Jev to allow or deny routine tool calls in auto mode, and hands uncertain ones to the built-in classifier

Claude assigns you the tasks only you can do. They wait in a My tasks pane, counted under the prompt, until you answer or reject them, so nothing Claude needs from you gets lost in the chat.

A pixel Pokémon that lives at the right edge of the band above the prompt

Read-only probe: lists system prompt sections and reminder types into /tmp/roof-probe-log.json

Replace Claude Code's built-in reminders with your own words: silence four nags, rename 'the user' in template sentences, drop the CLAUDE.md preamble.

Diagnoses wrong processes, workflows and habits in Claude Code sessions: fires on plan acceptance, workflow launches, pace complaints and the clock, names the lean alternative with its measured cost, and lets the user fix it in one click.

Host-bound cognitive lifecycle events for The Homie.

Measures a Claude Code session's fixed startup context payload per item at a pinned binary version, including per-tool attribution of the built-in tool pools that /context reports only as totals, derived by A/B deny differencing. Falls back from an SDK meter to a headless /context parser, then to an error rather than a wrong number. Keeps a before/after ledger per lever. Shows a live context breakdown in a pane or above the prompt. Report-only by default; fix applies one approved project trim.

Per-session context-window observability. A mod writes each session's context-window figures to a snapshot file and shows the person a toast and a transcript line when the context zone (smart, acceptable or dumb, from bands in zones.json) worsens; with model_notices on (off by default) it also tells Claude the zone at each crossing, approach and compaction. /context-guard prints status and toggles an optional band row. It answers a status tool. A PostCompact hook leaves a marker.

A mod that masks email addresses and secret-shaped text in what Claude Code draws, for recordings and screen shares. It changes the screen only; the transcript, the model's context and the files on disk are untouched, so it is not a privacy or data control.

review UI rewrite

Claude Code operations: audit-skill-visibility (skills the listing budget hides), inventory (every invocable surface), audit-install-state, audit-performance (slowness), audit-mod-posture (mods and the guard), audit-native-overlap (skills duplicating built-ins), observability (telemetry), known-issues, changelog, behavior-probes, prerequisites, check, machine-profile, plugins (update the fleet), morning-brief, lanes (background loop sessions), setup. Plus opt-in hook event logs.

Shared rate-limit guard for loop lanes and Claude: a hooks module writes the subscription rate-limit windows to a fixed machine-scope file and toasts window changes (telling Claude too only with model_notices on, off by default), a StopFailure hook records rate-limit stops reactively, and a reader contract fixes how consuming sessions pause and resume.

Hands the small judgments in a session to TypeSafe's Jev, a System One model that returns typed answers instead of generating text: what kind of request each prompt is, which model tier a subagent needs, whether an edit breaks a rule already written in your CLAUDE.md or AGENTS.md, and which transcript blocks survive compaction. Kept blocks carry over verbatim: no LLM writes a summary.

marketplace fails

Know your Claude limits before they hit you (Live band above the Claude Code prompt): 5 Hour and Weekly usage (whole account) with a live usage forecast against limit resets, context window usage, one-click Compact and per-session Auto compact (30% default).

Experimental Claude Code mod for Phase Runner: a pane with phase-viewer's screens and the live run, a toast when a run needs you, and /phase-status with no model call.

review UI rewrite

One loop: work: for work worth doing properly: clarify, an approved plan that is the run's only authority, delegated implementation, independent verification against computed gates, and human review. Domain workflows for code (test-first), data, writing and workshop presentations.

review UI rewrite

提问弹窗标题：在提问弹窗顶部显示消息编号和问题数，选项不变 / Question header: shows the message number and question count above the question dialog; options unchanged

助手：只读子 agent（Haiku），写三行周报 · 点按钮才派；「Claude 可派助手」默认关闭，在「安全」页开启 / Assistant: a read-only subagent (Haiku) that writes a three-line report · sent only when you press its button; "Claude can send the assistant" is off by default, switch on in Safety

核心：风格、中英文界面、消息编号和 /lemo-mod 面板，以及给其他 lemo mod 用的 $.lemo / Core: styles, zh/en interface, message numbers and the /lemo-mod panel, plus the $.lemo helpers other lemo mods use

限时：Claude 单轮工作超时自动中止 · 等你确认的时间不计 · 默认关闭，在「安全」页开启 / Time limit: stops Claude when a turn runs over time · time waiting for you does not count · off by default, switch on in Safety

记录：面板显示每轮统计 · 日志和自动分类默认关闭，在「安全」页开启 / Journal: per-turn stats in the panel · Log and Auto tags are off by default, switch on in Safety

抽签：给 Claude 增加抽签工具 draw_lot，抽到的签显示为签文卡片 · 默认关闭，在「安全」页开启 / Lots: adds a draw_lot tool for Claude and shows drawn lots as fortune cards · off by default, switch on in Safety

用量横条：输入框上方显示上下文、5 小时额度、花费和 git 分支；状态栏显示消息编号（已有自己的状态栏时默认关闭）；点「检查」才联网查新版本 / Usage meter: context, 5-hour quota, cost and git branch above the prompt; a status line with the message number (off by default if you have your own); checks for updates online only when you press Check

番茄钟：专注计时，到点在横条提醒并询问下一步 · 点了才开始 / Focus timer: times your focus, then a reminder on the band and a question about what next · starts only when you press it

会话小结：把本次会话汇总成三行摘要 · 点了才生成，对话越长用量越多 / Session recap: summarizes this session in three lines · runs only when you press it; longer conversations use more

换皮：消息加编号 T01、T02，回复标注「回复 T03」，工具行和耗时行使用风格样式 / Skin: messages numbered T01, T02, replies marked "Reply T03", tool rows and turn lines use the style

音效：调用工具、一轮结束、限时中止时播放提示音 · 默认关闭，在「安全」页开启 / Sounds: plays sounds on tool calls, turn ends and time-limit stops · off by default, switch on in Safety

加载词：工作时的加载词换成风格词并附上消息编号 · 已自定义转圈文字时默认关闭 / Spinner words: spinner words while Claude works come from the style, with the message number · off by default if you set your own

笔记：在面板记下笔记 · 压缩后自动摘要（Haiku）默认关闭，在「安全」页开启 / Notes: keep notes in the panel · Auto summary after compaction (Haiku) is off by default, switch on in Safety

说话方式：简短模式（每次回答不超过三句）、记录口吻（首行说明看了什么，末行给出结论） · 默认关闭，在「安全」页开启 / Tone: Brief mode (three sentences or fewer) and Notebook voice (opens with what it looked at, closes with a conclusion) · off by default, switch on in Safety

朗读：超过 30 秒的任务完成时，朗读编号和用时 · 默认关闭，在「安全」页开启 / Speech: reads out the message number and time when a task over 30 s finishes · off by default, switch on in Safety

定时提醒：N 秒后代你给 Claude 发消息，请它汇报进度 · 点了才开始 / Reminder: sends Claude a message for you after N seconds, asking for a progress update · starts only when you press it

Your AI B2B SaaS marketing team: 90 specialist agents in 19 skills across content and editorial operations, SEO/AEO, paid media, social, email, design, sales enablement, product marketing, integrated demand-gen campaigns and channel planning, founder-led GTM and public-sector marketing, account-based marketing (ABM), growth/PLG, analytics, comms and analyst relations, partnerships, events, developer marketing, project management, and client ops, coordinated by the CATALYST multi-agent orchestrator.

pi agent for claude: a `pi` subagent type whose model steps run the pi CLI, so pi subagents and agent-team teammates live in Claude Code's own agent UI, streamed live.

An extra-mods menu in the bottom corner: Zen mode with a progress band, Claude Code themes from terminalcolors.com, model, effort and permission mode controls and a status line.

review UI rewrite

Xcode in Claude Code: build, run and tests pane, scheme/destination band, notifications and SwiftUI previews over headless Xcode MCP

While Claude works, Clawd acts out the spinner's -ing word above the prompt, and the tool that is running

Cheap, fast, typed AI judgments inside Claude Code via the jevctl CLI (verify, screen, find, classify, extract, rerank, match, route, ask, batch), plus verbatim Jev-guided context compaction that replaces the built-in summary.

Shared Claude runtime adapter: normalises function-hook events into ClaudeRuntimeEvent, exposes $.runtime (emit bus, supports, judge, snapshot) to other plugins, and writes a JSONL event log

marketplace fails

Harness Mod layer: heartbeat/canary, native usage, rewrite explanations, measured subagent accounting, supervisor routing, secret redaction. Modes per guard in ~/.claude/mods/mods-config.json (classic | shadow_mod | mod).

marketplace fails

REALITY BENDER: intercepts a harmless tool result and alters what Claude receives; prints REAL vs RECEIVED

BLACK BOX RECORDER: structured per-session timeline of prompts, model steps, tool calls, subagents and denials; /blackbox prints it

GUARDIAN: executable policy that protects lab/demos/protected/** from Write/Edit/Bash, whatever CLAUDE.md says

MATRIX: freezes a harmless tool call before it runs, shows what Claude intended, asks you, then continues or denies. /matrix on|off

MIDDLEWARE TUNNEL stage 2: Cost analyzer: estimates the call's token cost from the args size

MIDDLEWARE TUNNEL stage 1: Observer: sees the call first, stamps t0, prints what enters

MIDDLEWARE TUNNEL stage 3: Policy: allows everything harmless; denies a matching rule

MIDDLEWARE TUNNEL stage 4: Telemetry: records the call and its result for the ledger

Interactive UI probe: ui.render components, toast, status, log, pane, command register

Interception probe: deny, rewrite, replace result, delay, throw, register tool, classify

Permission override, subagent visibility, agent.spawn model rewrite, turn.step routing, prompt context injection

Spike: does $.session.compact() work from a plugin, and what does the session.compact event carry?

Spike: does a tool.describe rewrite reach the model, does agent.offer hide a type, does prompt.suggest fire?

Spike: safe-boundary handoff (bundle written at turn.complete from the runtime bus) and $.turn.abort semantics

CLAUDE RUNTIME X-RAY: wildcard inspector with filters (/xray all|tools|subagents|session|usage|ui|errors|unknown|off), a HUD above the prompt, and a black-box timeline file

COSTCLAW LIVE: per-request cost, cache decay and repeated-read detection while the session runs, and it serves the Nth identical Read of an unchanged file from its own cache instead of paying for it

PRODUCTION GUARD: ambient project/environment context + a transport-independent policy core (DashClaw evidence classifier + offlocal policy engine) enforced natively on tool.call and tool.check. Blocks, asks, or contains before the command runs. /prodguard

SUBAGENT SUPERVISOR: routing-and-budget Mod. Rewrites a spawn's model onto the capability graph instead of denying it, keeps a measured per-agent-type token ledger in $.store, caps tool calls per subagent, and shows live subagents + the last routing decision above the prompt. /supervisor prints the routing log and the ledger.

The agent flow pane: /flow opens a live tree of the session's subagents and teammates beside the transcript, fed by engine events, with a text fallback where no pane can be drawn.

/explain: get answers as ASD-STE100 prose, a diagram, an HTML page, or a 3b1b-style video

See what you paste into Claude Code: image thumbnails and long-text previews above the prompt instead of bare [Image #1] and [Pasted text #2] tags

A cockpit for Claude Code: plan bars and subagent strips, context, usage limits with reset countdowns and burn rates, spend, the model route and a pet. Six bar styles, three collapsed looks. CLI and Desktop.

Writes a cold-start handoff to handoff/ before the context fills up or compacts, and offers to resume from it in the next session.

A plain-English receipt after every turn: files created, changed and deleted, by tool or via shell, with failed attempts listed apart.

review UI rewrite

Context window and session cost meter, drawn above the prompt.

Coral, ink and cream reskin of Claude Code with risk-coded tool rows. Toggle with /skin.

See inside a turn: a live Gantt timeline of model requests, tool calls and subagents, with tokens and cost per lane.

Route supported subagent work to Sonnet or Haiku while keeping the main conversation on its model. /router for controls.

A side panel of the files this session created, with Open, Reveal in Finder and Copy path.

review UI rewrite

Watch where Claude looks: a live treemap of the repo that lights up as files are read, searched and edited.

review UI rewrite

Explicit bookmarks for Claude Code sessions you want to come back to. /bm saves one, /bm open lists them with a copyable resume command.

A pixel-art crab above the prompt that reacts live to what Claude is doing. /pet for commands.

GitHub Flavored Markdown in the transcript: alerts (> [!NOTE]), task lists, strikethrough and Mermaid diagrams

review UI rewrite

Gauges how long the conversation's prompt cache stays warm, its hit ratio and misses, and what the next message re-caches once it has gone cold

Shares notices with every session in the same repository, or on the machine

Watches the session's pull request and wakes the session when it is merged or Codex reviews it

review UI rewrite

Shows context, 5-hour, weekly and per-model weekly limit usage above the prompt

Lets Bash commands written for bash run under zsh: unmatched globs, leading =, and a missing timeout

One line above the prompt: context tokens with their weather, a bar per recent prompt, your 5-hour and 7-day limits against the time elapsed, the prompt cache with its time left, the session cost with the last prompt's share in dollars and in points of the 5-hour limit, and the subagents running. Labels in English, French or Russian. After Token Weather (Anthropic, Apache-2.0) and the ideas of usage-meter (HolyGrail) and prompt-cache-control (Daniel San).

Enhancements for Claude Code auto mode

GitHub workflow, Actions monitoring, and rulesets management

GitLab workflow best practices and glab CLI usage

Drive the herdr terminal workspace manager: panes, tabs, workspaces, and sibling coding agents

Keep Claude-driven browsers and builds from saturating the machine

Structured event log for mods: `$.modEvents.emit` writes per-session OTel log records as JSONL

Planning mode guidelines and context injection

Pick a reply's `! <command>` suggestion by click or digit to fill it into the shell prompt

Warn the model mid-turn as the 5-hour or 7-day limit nears

Quote the bash habits that misfire under zsh before the Bash tool runs them, and explain zsh errors after

Compact once after 50 idle minutes, while the 1h prompt cache is still warm.

marketplace fails

Status band above the prompt in the Claude Desktop app's Code tab: folder, git branch and state, session age, prompt count and cost, a context meter, 5-hour and weekly plan usage limit meters with reset countdowns, last-turn model, duration and cache hit rate, idle time against the prompt-cache TTL, compactions, server-side tool use, subagents by state, and one row per workflow run with live done/failed counts. A notification (or toast) at 80% and 95% of a usage limit. Draws only on the desktop surface, so a CLI statusLine is left alone.

Watch the agent work: every step in plain English, with a live time-saved counter

Race Claude Code sessions on a task: live split-screen race track scoreboard

Failing tests spawn a pixel boss; every run that fixes tests lands a hit, zero failures is a KO with loot

Live session cost odometer above the prompt: a burning fuse, real-world comparisons, threshold alerts

A pixel pet that lives in a pane and reacts to what Claude does: eats on tool calls, gets sick on failures, panics on rm -rf, sleeps when idle, evolves as you ship

Gmail + Slack alerts inside Claude Code: toasts, status count, and an Alerts pane

A pane of Claude's dry inner thoughts about your session, typed out live

Dangerous Bash commands need launch codes: red alert pane, siren, a code to arm and a LAUNCH to fire

Spotify Wrapped for a Claude Code session: an animated stat reveal and a shareable PNG card

Live TV play-by-play of your Claude Code session, spoken aloud with crowd effects

Mission control for subagents and agent teams: who is running, what each is doing, who spawned whom, who is talking to whom, and how they overlap

Route Claude's desktop computer use through Codex's computer-use engine, one Codex session per Claude session so threads can drive apps in parallel. Toggle with /codex-cu on|off.

A lead chat creates real Claude Code sessions on chosen models, watches them live and steers them.

Prayer times, the Hijri date, adhkar, a daily ayah, sunnah fasting, Ramadan, Jumu'ah, tasbih and a salah break in Claude Code, each switched on or off from one dashboard: /muslim.

A desktop-app look for Claude Code's terminal: colored shell commands in prose, a Bottom line card, content-fit code cards, borderless tables, gutter marks beside paragraphs that need you, painted tool rows and turn headers.

review UI rewrite

btop-style live dashboard for Claude Code internals: /cctop opens it in a side pane, cctop-insights lets the session answer questions from its numbers

Discord Rich Presence for Claude Code - Show your session on Discord

A progress bar of the current task above the spinner, in the terminal and the Desktop app

Agentic development workflow covering planning, analysis, implementation, and QA automation

A band above the prompt showing your 5-hour and 7-day limits, context window and prompt-cache hit rate, styled for the terminal and the desktop app

Keep the main conversation's prompt cache warm by automatically forking its last request, tools denied and billed like any request, shortly before the cache expires; switch its lifetime between 5m and 1h, and show each refresh's cost and estimated saving.

KOZMOS Almanac: what this session is made of. Slash commands run, skills invoked, subagent types spawned, MCP servers used, models seen, plugins whose tools were called and tool families, each with counts and first-use time, as ranked bars and tag clouds; plus the engine version and how many installed commands went unused. /almanac toggles it.

KOZMOS Aurora: northern lights above the prompt while the model thinks. Shifting curtains whose intensity and speed follow the effort level (low is gentle, max is wild), a calmer single wave while it responds, and a label like 'thinking · xhigh · 12s'. /aurora toggles it.

KOZMOS Ballast: a context coach. A toast when the context window passes warnAt, then a band with an anchor and a water-level gauge at actAt: "⚓ ctx 89%: compact before the next big task" with Compact now (queued to the end of a running turn) and Not now. Optional auto-compaction between turns, and a before→after toast for every compaction. /ballast toggles the band.

KOZMOS Blackbox: a flight recorder above the prompt. A live Gantt of the current turn: the main loop and each subagent on its own lane, every tool call a bar in its family color, model requests as ticks, a moving now cursor on an auto-scaling time axis; the last turn stays dimmed until the next. /blackbox toggles it.

KOZMOS Breadcrumbs: the session's web trail in a sidebar. WebSearch queries with result counts and top hits, WebFetch URLs with status and size, and browser-like MCP visits, grouped by domain with a monogram badge, clickable links and a copy-all-as-Markdown button. /breadcrumbs toggles it.

KOZMOS Bridge: a mission-control sidebar. Model and effort, context fill, 5-hour and weekly limits with reset countdowns, session cost and burn rate, git branch and tree state, CPU/RAM/GPU, live subagents and the tool running now. /bridge toggles it.

KOZMOS Chronicle: the session as a timeline. Prompts, turn ends with duration, tool count and spend, subagents spawned and finished, commits and pushes, failed tools, compactions, and rate-limit and context milestones, each stamped HH:MM:SS, newest first, with filters. /chronicle toggles the sidebar.

KOZMOS Clawdling: a pixel crab companion that lives above the prompt. It blinks and sways, scuttles while tools run, thinks in bubbles, sweats past 80% context, dozes near the 5-hour limit and dances on a five-turn streak. It earns XP from tools and turns, levels up into a party hat, sunglasses, a scarf and a crown, and quips about what it sees. /clawdling toggles it, /clawdling feed treats it.

KOZMOS Compass: the single most useful hint under the prompt. In order: the 5-hour limit near its end with its reset, a context window near full, agents still working, uncommitted changes after edits, a draft left waiting; otherwise the engine's own hint. /compass toggles it.

KOZMOS Epilogue: a recap band after every turn. Duration, tool families as colored chips, files written, subagents spawned, tokens and estimated cost, the session cost delta and failed tools, with an optional AI one-liner. /epilogue opens the last 20 recaps.

KOZMOS Faultline: the error lens in a sidebar. Every failed or denied tool call, refused or errored turn and API error, grouped by a masked signature with count, first and last seen, the agents that hit it, the full first error (expandable) and a copy button, under a severity strip of the last 10 minutes. /faultline toggles it.

KOZMOS Gearbox: tool analytics for the session. Per tool its calls, errors and error rate, average, p95 and max duration and total time as a sorted bar chart, the five slowest calls, calls per minute, and the main loop against subagents. /gearbox toggles the sidebar.

KOZMOS Gitscope: a git radar sidebar. Branch, upstream and ahead/behind, staged/changed/new/conflict counts and stashes, the last 8 commits on a graph rail, per-file +/- histograms from git diff, and a marker on every file this session edited. /gitscope toggles it.

KOZMOS Glyphfall: matrix rain of tool activity above the prompt while Claude works. Every tool call drops its name and detail as a wave of letters, colored by tool family, with bright heads and fading tails, over a sparse rain of katakana; the last five tools sit below as colored chips. /glyphfall toggles it.

KOZMOS Halo: a one-row mood line above the prompt whose color and motion tell what Claude is doing: calm blue breathing when idle, a violet flow while it requests and writes, a magenta shimmer while it thinks, the running tool's color racing across, a red flash when a tool fails, an amber blink when it waits for you. /halo toggles it.

KOZMOS Heartline: a neon EKG of the session drawn above the prompt. Every tool call is a spike colored by its family, every model request a blip, a failed tool a red inverted spike; the trace drifts green to red with context fill, with a live bpm (tool calls per minute) and ctx readout. /heartline toggles it.

KOZMOS Helm: one-press settings profiles in a sidebar. Seven built-in ways of working (Token saver, Fast flow, Max quality, Plan first, Guarded, Autopilot, Quiet focus) plus your own saved setups; each shows exactly what it changes, lets you leave any step out, applies /config rows the way the menu does, saves effort for the model in use at once, and every activation can be undone. /helm toggles it; <profile>, apply <profile>, save <name>, rm <name>, undo, list.

KOZMOS Hivemind: the agent swarm as a live tree. The main loop at the root and every subagent beneath its parent, each with a status glyph, type and task, model and effort, elapsed time, requests, tokens, estimated spend, context fill and the tool it runs now. /hivemind toggles the sidebar.

KOZMOS Hourglass: the limit oracle. For the 5-hour, weekly and spend windows: percent used, reset countdown, a burn rate learned across sessions, the time to 100 % and a verdict on whether the reset comes first. Toasts at 80 % and 95 %, an optional status line. /hourglass toggles the sidebar.

KOZMOS Hub: the launcher for the whole bundle. /kozmos opens a starfield sidebar listing every KOZMOS mod by where it draws, which ones are installed in this session, and a button that runs each mod's own command.

KOZMOS Launchpad: a pane of big launch buttons. It learns your most-used slash commands as you run them, keeps pinned favorites and saved prompts, and runs a command or fills the prompt box in one press. /launchpad toggles it; save <text>, rm <n>, pin <cmd>, unpin <cmd>, list.

KOZMOS Laurels: 32 achievements earned from what the session does (Hydra, Night Owl, Cache Wizard, Bug Squasher, Polyglot, Centurion, Committed and more), kept across sessions with their unlock dates. A toast and a six-second celebration band on each unlock; /laurels opens the medal gallery with progress toward the locked ones.

KOZMOS Lofi: generative lo-fi music while Claude works. Four original synthesized loops (focus, deep, night, sunny: warm chord pads, an electric piano, soft kick and hats with swing, vinyl crackle) loop from the start of a turn until it ends, and stop when Claude is idle. Opt-in: /lofi on. A tiny band shows '♪ lofi · deep' with a dancing equalizer while it plays.

KOZMOS Mantra: a themed spinner. The working line's word rotates through a pack of original gerunds (cosmic, zen, pirate, turkish, hacker) and its tail names the tool running now with its elapsed time, or the effort while the model thinks. /mantra <pack> switches, /mantra off hands the line back.

KOZMOS Marquee: a status-line ticker. Model and effort, context fill, 5-hour and weekly limits with the reset countdown, session cost, git branch with ahead and dirty counts, and the tool running now, refreshed every second; it scrolls like a marquee when longer than its width. /marquee toggles it.

KOZMOS Mosaic: a GitHub-style activity calendar across sessions. 26 weeks of days colored by turns, tool calls, tokens or dollars, with the current and longest streak, the best day, totals and today's line. /mosaic toggles the sidebar.

KOZMOS Orrery: subagents as planets circling a central sun above the prompt. Running agents orbit at a speed set by their tokens per second, finished ones settle into a faint outer ring, failed ones flash a red cross; a legend names up to four with type and elapsed time. /orrery toggles it.

KOZMOS Polaroid: /polaroid snapshots the session as a self-contained dark-neon HTML report in .kozmos/reports: model, duration, cost, tokens, context peak, rate limits, tool calls by family, a timeline of turns, subagents, files touched and commands run, all collected live. It answers with the path and copies it. /polaroid md writes Markdown instead.

KOZMOS Questline: task progress above the prompt. One segmented bar for the session's todos and tasks (done green, running violet and pulsing, waiting gray), the count, what is running now and for how long. Hides itself two minutes after everything is done. /questline toggles it.

KOZMOS Relay: the API latency lab in a sidebar. Every model request of the main loop and its subagents with latency, time to first token, model, effort, input/output/cache tokens, tokens per second and stop reason; a latency histogram, tokens/sec over time, model mix, p50/p95, the slowest five and cache-read share per request. /relay toggles it.

KOZMOS Resonance: a synthesized soundscape for the session. A soft tick per tool, a bonk on failures, a gong after long turns, a bell when a subagent finishes, an alarm when a rate-limit window crosses 80% or 95%, a chime on the first turn. /resonance mutes.

KOZMOS Skies: the session's weather above the prompt. Tool failures, context and rate-limit pressure, stalls and the local hour make clear skies, clouds, rain, storms with lightning, fog, a starry night, or a rainbow after a storm clears; an animated scene and a forecast line. /skies toggles it.

KOZMOS Spectra: the context X-ray. The context window broken down by category as a heat grid with a ranked legend, context growth per turn, and an estimate of how many turns remain before auto-compact. /spectra toggles the sidebar.

KOZMOS Stamp: a receipt on every finished turn. The closing 'Baked for 42s' line becomes a compact row of duration, tool calls by family as colored chips, tokens in and out, the turn's cost and the context fill with its change. /stamp toggles it.

KOZMOS Storyboard: Claude narrates its own progress. A chapter tool the model calls at each phase change drives a band above the prompt: the chapter title, a five-stage rail (explore, plan, build, verify, ship) with the active stage glowing, step 3/5 and a note. /storyboard toggles the band, /storyboard log opens the chapter history.

KOZMOS Switchboard: an MCP observatory sidebar. Every MCP server in the session with its tools, calls, errors, average and p95 latency, last call and a status lamp (healthy, erroring, idle, busy); servers with no calls yet too, and per-tool rows on expand. On the desktop a patch bay with glowing cables to the active servers. /switchboard toggles it.

KOZMOS Taskforge: a live kanban of the session's tasks. TodoWrite lists and TaskCreate/TaskUpdate tasks fall into Pending, In progress and Done columns with the active form, time in the column and the owning agent, under a progress bar. /taskforge toggles the sidebar.

KOZMOS Tempo: a Pomodoro focus timer above the prompt. A tomato, the phase, a ring (desktop) or a sizzling bar (terminal) and the time left; long breaks every fourth round, a toast at each phase end and a tally of focus rounds per day. /tempo start [work] [break], stop, skip.

KOZMOS Terminus: the session's shell history in a sidebar. Every Bash and PowerShell call of the main loop and its subagents with command, description, cwd, exit status, duration, output size and who ran it; filters for failed and slow runs, copy and re-run buttons. /terminus toggles it.

KOZMOS Thermal: a heatmap of the files this session touches. Reads, edits, writes and searches per file, a heat score that cools over time, grouped as a directory tree under the session folder: an indented heat tree on the terminal, a squarified treemap on the desktop, switchable by heat, reads or edits. /thermal toggles it.

KOZMOS Thrift: a frugal mode. /thrift on|off|auto; while on, every prompt carries a short note asking Claude to be economical (targeted reads, no broad searches or subagents unless needed, concise answers). Auto turns it on when the 5-hour or weekly limit passes autoAt and off after the window resets. A band with a spinning coin, and /thrift compares the average cost of thrift turns with normal ones.

KOZMOS Throttle: a car dashboard above the prompt. Speedometer of output tokens per second, fuel gauge of the 5-hour limit, temperature of the context, an odometer of session dollars, and check-engine, fuel and overheat lamps. /throttle toggles it.

KOZMOS Tidewater: the live diff of the current turn as tides above the prompt. Every file Claude edits gets its own pool of rolling waves, a green swell for added lines and a red ebb for removed ones, with file chips and totals like '+128 −42 · 6 files'. From the first edit of a turn until the next prompt. /tidewater toggles it.

review UI rewrite

KOZMOS Tokenomics: the cost lab. Session cost as the headline, an estimated split by model and by token type, cache hit ratio, burn per hour, cost per turn, a pace forecast and daily and weekly totals across sessions with a 14-day chart. /tokenomics toggles the sidebar.

KOZMOS Toolmarks: badges on tool rows. Each finished call gets a family-colored glyph and its duration at the row's end, Bash calls their exit status, failed calls a red mark, and a folded group its total time and failures. /toolmarks toggles them.

KOZMOS Verdict: test, build and lint results above the prompt. Spots the test runners, type checkers, linters and builds Claude runs (vitest, jest, pytest, go test, cargo, dotnet, mocha, bun, tsc, eslint, npm run build, maven, gradle…), parses pass/fail/skip and error counts, and shows chips like '✓ vitest 142 · ✗ jest 3 failed · ✓ tsc 0 errors' with a red pulse on failures and a sparkline of the last 10 runs. /verdict opens a pane of recent runs with the failing test names and first error lines.

review UI rewrite

KOZMOS Vitals: a btop-style machine monitor sidebar. CPU with a braille history graph, RAM bar and history, disk use, NVIDIA GPU load, memory and temperature, process count, and the Claude side: session time, tool calls, running agents. Samples every 2 s only while open. /vitals toggles it.

KOZMOS Warden: a guard for destructive shell commands. Before Claude or a subagent runs rm -rf on a broad path, git push --force, reset --hard, clean -fd, DROP TABLE, mkfs, format, curl | sh and their kin, warden asks you (Allow once / Deny) and refuses when no one can answer. A band flashes for 8 s after a block; /warden lists the rules and this session's log.

review UI rewrite

KOZMOS Warpdrive: a starfield at warp above the prompt. Stars stream out of the center at the speed of Claude's live output token rate: a slow drift while it thinks, hyperspace streaks while it writes. When a turn ends, a jump flash and a line like '⇢ arrived · 42s · 18.4k tok'. /warpdrive toggles it.

Animated pixel art above your Claude Code prompt that reacts while Claude works: seven scenes, or your own image or GIF. Zero tokens.

/park a session you will come back to; a new session in the same folder offers it back with one click.

多個 Claude Code session 分工時的派工看板：讀你寫的 mission 檔，顯示誰完成、誰在跑、誰在等你，狀態轉換時跳 toast

review UI rewrite

dispatch-board 的搭檔：worker session 把回報寫進結果檔後以一行結束，由 mod 在 turn 結束時把新增內容與最終回覆轉送給指揮站；背景工作進行中的回覆暫存、延後併送（mission 檔寫了 station 才啟用；需 Claude Code ≥ 2.1.296）

bg session 在 worktree 做完的分支，由你在面板按一顆按鈕快轉進 main：交接單只帶參數，步驟由 mod 推導，執行前檢查、執行後驗證

review UI rewrite

把 context window 畫成一顆腦：工具呼叫點亮對應腦區，用量越滿越擁擠，compact 時睡眠整理

Translates the Georgian in each prompt to English before Claude reads it, without echoing the translation

Fixes the grammar and spelling of each English prompt before Claude reads it, and shows what changed

nanoflow by nanosite.ai: your dev workflow, live. Worktrees on their own ports, tickets, PRs/CI, checks and Playwright as a dashboard pane, a progress status line and toasts, plus dev skills.

Ambient scenes above the prompt that move with your agent's work: a fire, bubbles, surf, ski, rockets, a hot-air balloon and more, by day or night. /flow to control it.

A floating status card above the Claude Code prompt: model, effort, context, 5h/week limits, cost, directory and branch, GitHub CI and deploys: with clickable settings and a progress-row API any script or mod can feed.

Shows Claude's questions in a side pane with context, option details, an AI explanation and answer history; also explains questions Claude asks in plain text.

TypeSafe's Jev decides how Claude Code works on each prompt: the main conversation's reasoning effort (low to xhigh, raised to max mid-turn when tool calls keep failing), each subagent's model, whether to split the work across subagents, and the one skill the prompt needs. Brings in custom OpenRouter models (a junior coder reviewed by Opus, or cheap bulk work) and Codex or OpenCode as reviewers, in modes you choose, each health-checked. Reads the prompt with recent context, records every decision, and suggests its own tuning. Backends: OpenRouter, TypeSafe's API, the Vercel AI Gateway, or Claude Code's built-in classifier. Built on claude-code-templates' jev-model-router and jev-skill-suggestion.

Live view of a running Workflow: each stage as it starts and finishes, with what it is doing right now, in a pane, a band above the prompt and the status line.

像素剩余额度、固定小房间里的十二种状态动画与浅深主题，配合主会话优先、后台等待、真实压缩与实际阈值读数

Generate and edit images using Google Gemini, OpenAI GPT Image, xAI Grok Image, and OpenRouter APIs

A one-glance summary of jev-blindspot in the band above the prompt: the blind spot count, the deeper question, and the top findings for the prompt you just sent.

Side pane listing unfinished and in-progress todos from all Claude Code sessions, grouped by project

5h / 7d usage, token totals and cost pills above the prompt

Read ツールの結果に含まれる認証情報の形をした文字列を、モデルが読む前に伏せる。

Replaces Claude Code's compaction with a structured handoff: a fork of the session writes it from the prompt cache, and the conversation continues from exactly that handoff.

A pane listing this session's subagents (Agent tool and workflow agents) with their status; select one to see its prompt, text, tool calls and results.

review UI rewrite

Click a file path under a Read, Edit or Write row to show that file's contents in a pane beside the transcript.

review UI rewrite

A pane listing the project root repository's open pull requests with titles; click one to open it in the browser.

A side pane for Claude Code that lists the subagents a session runs, what each is doing and what it costs in tokens, with each one's conversation a click away.

awsh for Claude Code: an `aw` subagent type whose model steps are answered by a session on the local awsh harness daemon, so any harness and any backend the daemon can drive runs as a native Claude Code subagent, streamed live. Also carries the Aither World skills, the `aither` output style, MCP servers for installed bricks and portable brick hooks.

Compacts a long tool result through the AitherOS decision door before the model reads it: per LINE SHAPE, learned from outcomes, never dropping a traceback or a summary line.

After an edit that adds a GitHub Actions step pinned to a tag or a branch, names the commit SHA to write instead, from the GitHub API.

Picks the recommended option of a question that waited unanswered for a set time, so a session you left does not stop. Off until you turn it on.

Shrinks each Bash result before the model reads it: per-command filters for git, test runners, linters, package managers, containers and file tools, the full output kept in a file.

Shows the background shell tasks of the session on the status line or in the shared sidebar, with the oldest one's age, and stops one from the /bg-tasks pane.

Runs proof-driven bug hunt rounds with /bughunt: each round proves one bug with a failing command the mod runs itself, fixes it, and proves the fix, and the loop stops on a blocked or unverified round. /bughunt collab runs a read-only scanner, planner and critic in sequence.

Keeps the 1-hour prompt cache warm for a window you set, or with no end in every session under always, with one cache-shared fork per idle stretch and one keep-warm message after a resume, arms a window for you after a paid cold write, and shows the cache state, the cold price and this session's cold writes.

Names the uncommitted files to you at the end of each turn, and tells the model at the next prompt to commit each finished piece as it lands.

Parses each JSON, YAML, TOML or .env file an Edit or Write touches, notes a parse error at once instead of at the next build, and says so again when a later edit fixes it.

Hands a skill or command call the current text of its file when the file changed on disk after the session loaded it, and hands the model a rules file or the global CLAUDE.md that changed on disk.

After the model changes a function signature with Edit, adds the callers ripwire finds to the Edit's result, so the model fixes them before the build does.

Asks a council of models about a hard problem, when the model calls its tool or you run /council <question>: Claude models, and Gemini models when gemini-core has a key, answer in parallel, and the model of the session writes one verdict from their answers.

Checks each package the model installs against its registry and OSV.dev, and stops a missing, brand-new, look-alike, outdated or vulnerable one, naming the latest version.

Sends a desktop notification when a question or a plan waits for your answer, and when a turn ends or fails.