Auto Approve Claude Plan — Development skill for Claude Code
auto-approve Claude Code plans instantly — optional Craft.do archiving.
How to install Auto Approve Claude Plan
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open yigitkonur/hooks-claude-code and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Auto Approve Claude Plan does
auto-approve Claude Code plans instantly — optional Craft.do archiving.
Alternatives in Development
- Implementation Plan — You are tasked with creating detailed implementation plans through an interactive, iterative process 10k ★
- Dev Plan — /dev-plan — Génère les plans techniques d'1 FEAT sans coder 193 ★
- YOLO/Auto-Mode 분류기 시스템 — 관찰 위치: Claude Code 내부 아키텍처 Tool 호출을 자동 승인할지 차단할지 결정하는 2단계 보안 분류 시스템입니다 149 ★
README
auto-approves Claude Code's "ready to code?" plan dialog so you stop clicking a button 50 times a day. an optional **Deep Plan** mode also nudges Claude to reason harder *while you're in plan mode* and gives you effort-tier launch aliases. pure bash, no dependencies beyond `jq`.
bash <(curl -fsSL https://raw.githubusercontent.com/yigitkonur/auto-approve-claude-plan/main/install.sh)
[](https://www.gnu.org/software/bash/) [](#) [](https://opensource.org/licenses/MIT)
the problem
Claude Code has a plan mode. when Claude finishes writing a plan and calls `ExitPlanMode`, it fires a `PermissionRequest` event and waits for you to click approve. every single time. this hooks into that event and returns `{"behavior":"allow"}` immediately — and also asks Claude Code to land the session in `bypassPermissions` so subsequent tool calls don't prompt either.
two modes
the installer asks you to pick one:
| mode | what it does |
|---|---|
| 1 — Classic (default bypass) | approves every plan instantly and lands the session in bypassPermissions. nothing else is touched. install and forget |
| 2 — Deep Plan (recommended) | everything in Classic, plus a prompt-level boost: while you are in plan mode every prompt is nudged to reason deeply (auto-reverts when you leave plan mode), effortLevel is set to low as a cheap normal-mode baseline, and two launch aliases (claude-plan / claude-fast) are added to your shell |
The old **Orchestrator** mode has been removed. Re-running the installer over an existing Orchestrator install cleans up its `PostToolUse` hook automatically.
what Deep Plan actually does (and a candid scope note)
Deep Plan is honestly a little **outside the ori
Related Skills
Auto Trigger
After any recon/probe step, route observed signals to the right tools and fire the auto-plan (Balanced policy)
Plan Auto Context Enhanced
Create implementation plans with automatic context selection. Just describe what you want to build in natural
Gdrive Push
Push this course's courseware (PPT + PDF, Learner Guide, Lesson Plan, assessments) to the user-provided Google
Lightclaw
Control local Codex and Claude coding agents from your phone via Telegram — review plans, approve bounded work
TokenLink
Native macOS menu-bar monitor for Codex, Claude, Kimi, MiniMax, and GLM coding plans, with optional M5Stack St
Creative Writing
Workshop-voiced craft pack for prose narrative — fiction and creative nonfiction. Three explicit modes (draft
Related Agents
Part Implementer
Craft implement phase worker. Executes exactly one plan part via strict TDD and lands it as one atomic convent
Auto Runner
Sonnet worker for the trivial-PR auto-approve path (docs-only / dependency patch bumps that passed trivial-che
Adal CLI
AdaL CLI (SylphAI, named after Ada Lovelace) is a terminal coding agent built for developers who want to see a