Инструменты

Persistent memory across sessions: hooks record tool calls and generate semantic summaries, then inject relevant context into new sessions, with history search and a web viewer. The hosted memory service needs sign-in, or use your own Anthropic plan or model key.

Four guidelines based on Karpathy's observations of common LLM coding mistakes: think before coding, prefer simplicity, make only necessary changes, and work toward verifiable goals. Ships as a plugin skill or works as a CLAUDE.md. Community-made, not by Karpathy.

'Laziest senior engineer' mode: checks whether the work is needed and whether existing code, the standard library or platform features cover it, then writes the minimum code while keeping validation, error handling and security. For Claude Code and Codex; its hook needs Node.js.

Generates hook rules from a one-line description or by analyzing recent conversations. Rules are Markdown files with match conditions that block or warn on dangerous commands and unwanted patterns, and take effect without a restart. Official Anthropic plugin.

Makes the agent answer in terse 'caveman' style to cut output tokens, compressing only prose while leaving code, commands and errors intact. Includes a local proxy that compresses logs, test output and other input. Works with Claude Code, Codex and 30+ other agents.

Addy Osmani's production-grade engineering skill pack covering spec, planning, implementation, testing, review and shipping, with commands like /spec, /plan, /build, /test, /review and /ship. Supports both Claude Code and Codex.

Orchestrates multi-agent teams using Claude Code's experimental Agent Teams feature: parallel multi-angle code review, competing-hypothesis debugging and feature work split by file ownership. Enable CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS first.

Context-driven development plugin: first captures product vision, tech stack and workflow docs, then generates a spec and phased plan for each feature, executes with TDD and supports per-task rollback. Ported from Google's Conductor for Gemini CLI.

Audits whether CLAUDE.md matches the current codebase and scores its quality, and /revise-claude-md writes lessons from the current session back into project memory. Official Anthropic plugin.

Implements the Ralph Wiggum loop with a Stop hook: feeds Claude the same prompt repeatedly so it keeps iterating on its previous file changes until it outputs a completion marker or hits the iteration limit. Official Anthropic plugin.

Scans your codebase and recommends the best hooks, skills, MCP servers, subagents and slash commands by category. Analysis only, no file changes; a quick way to set up Claude Code for a project. Official Anthropic plugin.

Anthropic's official plugin marketplace, built into Claude Code by default, with 300+ official and partner plugins for development, security, design and office collaboration. Browse and install with /plugin.

Installs ready-made subagents, slash commands, hooks, MCP configs and skills for Claude Code with one npx command, browsable online at aitmpl.com. Also includes session analytics, health checks and other helpers.

Processes large tool outputs in a sandbox and returns only the results to context; logs session events in SQLite with FTS5 so work continues after context compaction. Supports Claude Code, Codex and a dozen other clients. ELv2 licensed.

Status line plugin for Claude Code showing the model, Git branch, context usage and rate limits live below the input, optionally with running tools, active subagents and todo progress.

Parses a PRD into a dependency-aware task list so Claude Code, Cursor and other agents pick up the next task in order, split subtasks and update progress. Offers both a CLI and an MCP server. Needs at least one model API key, or reuse a Claude Code or Codex login.

Parses a PRD into a dependency-aware task list with complexity analysis, task breakdown and research; the plugin bundles slash commands, subagents and the MCP server. Needs a model API key, or reuse a Claude Code or Codex CLI login.

160+ specialized Claude Code subagents, packaged as plugins in 10 categories such as core development, languages, infrastructure, testing and security, and data and AI, so you install only what you need.

Adds a structured dev process to Claude Code: 30 /sc slash commands for brainstorming, implementation, testing, troubleshooting and research, 20 domain expert subagents and 7 behavior modes, with optional MCPs like Context7. Installs via pipx; a marketplace version is in progress.

Pressures the agent with big-tech PUA and PIP-style rhetoric plus structured debugging steps and initiative checks, pushing it to change approach after failures and show evidence before claiming done. Modes include always-on, encouragement and P9 lead. Works with Claude Code, Codex and more.

Reports token usage and cost for Claude Code, Codex, Gemini CLI and other coding agents from local logs, by day, week, month or session, and shows Claude Code's 5-hour billing windows.

Customizes the Claude Code status line to show model, Git branch, token and context usage, session and weekly limits at the bottom of the terminal. Configured through an interactive TUI with Powerline styling. Claude Code only.

Reads local session files to report token usage and cost by project, model, task and branch for Claude Code, Codex, Cursor and dozens of other coding agents, no sign-up needed. Also offers config optimization tips, quota checks and spend-cap hooks. Needs Node.js 22.13+.

Spec-driven development and context engineering framework (formerly Get Shit Done): a five-step loop of discuss, plan, execute, verify and ship, handing heavy work to subagents with fresh context to avoid degradation in long sessions. Supports Claude Code, Codex and more.