Claude Code Tools
Practical productivity tools, plugins, hooks, and skills for Claude Code, Codex-CLI, and similar CLI coding agents.
- Source repo
- pchalasani/claude-code-tools
- Stars
- ★ 2k
- Last updated
- 6d ago
- License
- MIT
- Primary language
- Python
- FA score
- 66/100 · Some gaps
At a glance
- How it runs
- Works with
- Portable with changesCodex · Claude Code
- Cost
- Free software; you pay for model usage
- Setup effort
- Medium · a few setup steps
- You'll need
- Typical use
- A Claude Code developer who wants safety hooks to prevent the agent from running dangerous commands or leaking environment variables (safety-hooks, sasy-guard, env-safe).
- Not a fit if
- Users unwilling to install tmux or a Rust toolchain
- Teams that want setup fully documented in the README, not an external docs site
- Maintenance
- Commits in the last 30 dayslast commit 2026-11-10latest release voxtype-v0.3.19 (2026-11-10)+0.9 ★/day over 7 days
- Source review
- 66/100 · Some gaps
What does this agent do, and when should you use it?
pchalasani/claude-code-tools is a collection of practical productivity tools for CLI coding agents such as Claude Code and Codex-CLI. It ships a broad set of standalone CLI tools — including aichat, voxtype, tmux-cli, amux, agent-tunnel, lmsh, vault, env-safe, sasy-guard, statusline, fix-session, github-wake, and msg — plus skills, agents, hooks, and plugins. The core package is published on PyPI as claude-code-tools, and the aichat-search component is published on crates.io; the project is MIT-licensed. Full documentation lives at pchalalasani.github.io/claude-code-tools, covering installation, every tool, plugins, and guides. The repository is listed in Awesome Codex CLI and suits developers who want to make terminal-based coding agents safer and more efficient.
The repository provides installable CLI tools and Claude Code plugins: aichat offers chat-based interaction backed by aichat-search (Rust, on crates.io); voxtype provides voice input; tmux-cli and amux manage tmux sessions and multiple agents; agent-tunnel provides tunneling; vault and env-safe handle secrets and environment-variable safety; safety-hooks and sasy-guard provide safety hook protection; statusline renders a status bar; fix-session repairs sessions; the session-porting tool moves sessions between Claude and Codex; github-watch wakes an agent on a GitHub comment; msg handles inter-agent communication; there are also Google Docs and Google Sheets integrations and alternate LLM provider configuration. Everything is deployed and invoked via CLI commands or the Claude Code plugin mechanism.
- A Claude Code developer who wants safety hooks to prevent the agent from running dangerous commands or leaking environment variables (safety-hooks, sasy-guard, env-safe).
- An engineer using both Claude Code and Codex-CLI who needs to port session context between the two (session porting tool).
- A team running multiple coding agents in parallel that wants tmux-based session management (tmux-cli, amux).
- A developer who prefers dictating instructions to a coding agent by voice (voxtype, voice plugin).
- A maintainer who wants an agent to watch a GitHub repository and wake up when a comment arrives (github-watch).
- A user who wants an agent to read and write Google Docs and Google Sheets (gdocs, gsheets integrations).
How do you install or deploy this agent?
The README offers limited copyable install instructions: the core package is distributed via PyPI (pip install claude-code-tools should work; the badge points to pypi.org/project/claude-code-tools), and aichat-search is distributed via crates.io (cargo install aichat-search). Per-tool install commands, system dependencies, and credential setup are not given in the README — consult the quick-start page at https://pchalasani.github.io/claude-code-tools/getting-started/.
How do you use this agent?
After installation, each tool runs as a standalone CLI command (aichat, voxtype, tmux-cli, amux, vault, env-safe, statusline, fix-session, msg, github-watch), while plugin features (safety-hooks, voice, visual-brief) are loaded through the Claude Code plugin mechanism. The exact first working commands and flags are not listed in the README; refer to the per-tool documentation pages (e.g. /tools/aichat/, /tools/tmux-cli/) on the docs site.
What are this agent's strengths and limitations?
- Breadth of coverage: safety hooks, secrets management, session repair, and inter-agent communication address multiple pain points of coding-agent workflows in one repository.
- Supports both Claude Code and Codex-CLI, including session porting between them, reducing vendor lock-in.
- Legitimate distribution channels: a PyPI package, a crates.io component, and a dedicated documentation site; listed in Awesome Codex CLI.
- MIT license allows free use and modification.
- The large number of tools (over a dozen) means meaningful evaluation and integration effort per tool.
- The README is a card-based navigation hub with few inline details; users must jump to an external docs site to get started.
- Some features have specific runtime dependencies: tmux tools require tmux, and aichat-search requires a Rust toolchain.
- Google Docs/Sheets integrations and alternate LLM provider support require additional credentials and configuration whose constraints are only visible in the docs.
How does this agent compare with similar options?
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| Claude Code Tools This agent | 66 · Some gaps | CLIFree + model costs | ★ 2k | 6d ago | Python | Codex · Claude Code |
| CCB Multi-Agent Terminal Workspace | 79 · Good | CLIFree + model costs | ★ 3.6k | 6d ago | Python | Codex · Claude Code |
| Claude Council | 73 · Some gaps | Agent plugin / skillFree + model costs | ★ 943 | today | Shell | Claude Code · OpenAI API · Claude API |
| Oh My OpenAgent | 64 · Some gaps | Agent plugin / skillFree + model costs | ★ 71k | today | TypeScript | Claude Code |
How does FollowAgents rate this agent?
Why each dimension lost points
The sample shows capabilities that rely on tmux, Discord, hosted Jev, or CLI subprocesses; `agent-tunnel` also exposes explicit write, bash, and full-access options. This supports data-flow visibility and some sensitive-data handling, but evidence for default permissions, operation confirmation, external side effects, and recovery is limited, so those criteria are not scored highly. Dependencies have minimum versions, but the sample shows no security audit or lockfile evidence. The MIT license and author attribution are clear; publisher identity is unverified and earns no credit by itself.
The CLIs provide readable or structured failure paths for missing dependencies, absent TTYs or sessions, invalid responses, and malformed configuration; the Jev skill also sets bounded retries and stop conditions. Multiple tools, platforms, and external services are dependencies, while the sample gives little evidence about availability or fallback coverage. The README's Node.js requirement does not fully align with the Node.js version mentioned in the session-search skill, so consistency receives a moderate score.
The materials cover Claude Code, Codex CLI, tmux, and varied workflows. Visual Brief and Jev state clear use cases and skip conditions. The sample also identifies platform-specific features, authentication, and external services, but it does not define the boundaries and environment compatibility of the entire suite; trigger rules are only visible for the sampled skills.
The README directs users to detailed documentation and organizes feature entry points; SKILL files include names, trigger descriptions, workflows, and examples. The MIT license, package version, and release workflow are visible. Installation guidance covers basic skill installation and some CLI prerequisites, but package names and update paths for several capabilities are incomplete. The sample does not show a clear maintenance commitment.
The outputs are designed for practical use: Visual Brief defines a structured report, the Jev workflow constrains revisions, and tools provide JSON, tables, and clear CLI feedback. Session search, agent selection, and inter-agent communication offer concrete workflow value. Multiple components, external authentication, and hosted services add user cost, so cost-benefit is not full marks.
Key usage and boundaries can be traced to skill instructions, CLI options, and package configuration. Sample tests provide static corroboration for auth isolation, HTTP error handling, and input limits, but do not justify claiming execution succeeded. The README, manifest, code, and tests partly corroborate one another, with incomplete coverage. Skills distinguish verified, agent-reported, and unverified claims, supporting fact/inference separation, though the suite has no single evidence format.
- This is a static review of sampled files; unseen files are unknown, and no tools or tests were executed.
- Before using features that depend on Discord, tmux, credentials, or hosted Jev, check their environment requirements and what data they send to external services.
Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
voxtype-v0.3.19 voxtype v0.3.19
-
voxtype-v0.3.18 voxtype v0.3.18
-
Release notes excerpt
aichat trim --keep-last N: keep the tail of a session intact Deterministic trim now takes --keep-last N / -k N, which preserves the last N user turns verbatim while earlier history is trimmed as before. A turn is a user prompt plus everything that follows it — assistant text, thinking blocks, tool calls and tool results — up to the next prompt, so you keep whole, coherent exchanges instead of an arbitrary count of records that can cut a tool call away from its result. - Works across Claude trim, Codex trim, and Claude trim-in-place. …View on GitHub ↗
FAQ
Is Claude Code Tools free?
Is Claude Code Tools open source?
Is Claude Code Tools safe to use?
Is Claude Code Tools still maintained?
Who makes Claude Code Tools?
Thanks — we use this to improve the page.