Claude Code Permissions Generator

Pick a safe starting mode and your OS, and get a validated allow/deny permissions block for settings.json in under a minute.

Question 1 of 5

Pick a starting posture

Every option includes the same hard-coded guardrails below - project history, secrets, and the delete-everything command stay protected no matter what you pick here.

5 quick questions - free, no sign-up, nothing leaves your browser

FAQ

What does this permissions generator actually output?

A permissions block for Claude Code's settings.json: an allow array, only if you picked something to pre-approve, and a deny array that always includes the baseline guardrails - blocking force-push, reset --hard, rm -rf, and reads of .env and secrets/ - plus anything else you added.

Does allow beat deny, or the other way around?

Deny always wins. Claude Code evaluates rules in a fixed order - deny, then ask, then allow - and the first match by rule type decides the outcome, regardless of which rule is more specific. A broad deny rule can block a call that a narrower allow rule also matches.

Why doesn't this generator have an allow rule for git status or git diff?

Because it would do nothing. Claude Code already runs a built-in set of read-only commands, including git status, git diff, and git log, without a permission prompt in every mode. Adding an allow rule for them is a no-op some generators ship anyway.

How does the Windows path conversion work?

Claude Code normalizes every path to POSIX form before matching, so C:\Users\you\.aws becomes //c/Users/you/.aws internally. Type a Windows path with backslashes into the last question and this generator converts it for you; pasting the raw backslash form into settings.json by hand would silently match nothing.

Where do I put the generated permissions block?

Inside a permissions key in ~/.claude/settings.json for every project on your machine, or .claude/settings.json for one project you want your team to share. If the file already has a permissions key, merge the arrays in rather than replacing them - the paste-prompt output does this for you.

Is this permissions generator free?

Yes - free, no sign-up, and nothing you pick or type leaves your browser; the JSON is assembled locally. ClockedCode separately sells a complete done-for-you Claude Code setup, but the generator itself is not gated.

This generator builds the one part of settings.json most people get wrong by hand: the permissions block that decides what Claude Code can run without asking, and what it can never touch. Five questions, all click-only except one optional path field, and the last screen pairs the JSON with a plain-English readout of what the rules actually do.

Deny always beats allow in Claude Code, even when the matching allow rule is more specific - a broad Bash(aws *) deny still blocks Bash(aws s3 ls). And read-only git commands like status, diff, and log already run without a prompt in every mode, since they're on Claude Code's own built-in read-only list, so an allow rule for them does nothing. Both facts are easy to miss reading the docs cold, and both changed what this generator ended up putting in the output.

Pick your OS and the last question - protecting an extra path like your AWS credentials or SSH keys - gets the syntax Claude Code actually matches against. Type a Windows path with backslashes and the generator converts it to the //c/... form Claude Code normalizes to before matching; the backslash version you'd paste in by hand matches nothing. It also only ever writes one Read() rule per path, because a Read deny rule already blocks Edit on that same path as of Claude Code 2.1.208 - a second rule would just be redundant.

The guardrails - blocking force-push, git reset --hard, rm -rf, and reads of .env and secrets/ - aren't a checkbox here. Every starting mode ships with them, so there's no screen where an unchecked box quietly removes your safety net. What changes between Balanced, Cautious, and Workflow-friendly is only how much gets pre-approved on top: Cautious skips the pre-approval screen entirely, Workflow-friendly starts it pre-checked for tests, lint, and GitHub reads.

This generator only covers permissions. For hooks, environment variables, and model selection in the same file, the settings.json generator builds the rest of it, and the complete permissions guide walks through every mode and rule type this tool encodes, if you want the reasoning behind a specific answer.

I run Claude Code daily and this generator writes the same permission rules my own settings.json uses. If you want the finished version of the whole idea - a tuned settings.json, a tuned CLAUDE.md, and the full vetted tool stack, installed in one paste - that is what ClockedCode is. The link is at the bottom of this page.

Neo ZinoBy Neo Zino - builder of ClockedCode

Other free tools

Config generator

Output Style Generator

Pick a persona and a tone, and get a valid output-style file with correct frontmatter, ready to save and switch to.

Try tool
Config generator

Subagent Generator

Describe the job in plain English, pick a specialty, and get a complete subagent file ready to save.

Try tool
Config generator

MCP Config Generator

Pick the MCP servers you use and get the exact claude mcp add commands or a ready .mcp.json block, scoped and copy-pasteable.

Try tool
Config generator

Slash Command Generator

Answer 7 quick questions and get a custom Claude Code slash command with valid frontmatter, ready to drop into .claude/commands/.

Try tool
Config generator

Statusline Generator

Toggle the segments you want, watch the exact statusline render live, then copy the tested script and settings.json block.

Try tool
Config generator

settings.json Generator

Pick permission rules, hook presets, env vars, and a model override one step at a time, then copy or download the finished settings.json.

Try tool
Config generator

CLAUDE.md Generator

Answer 7 quick questions, get a tuned CLAUDE.md for your project in 30 seconds.

Try tool
Template library

Claude Code Templates

A curated library of copy-paste Claude Code templates: subagents, hooks, slash commands, settings, and CLAUDE.md starters, each one explained and pulled from a setup that gets used every day.

Try tool
Token scanner

UsageCut

Scan your Claude Code setup in 30 seconds and see exactly what is burning your tokens - bloated context, heavy MCP servers, and settings quietly draining your usage limits.

Try tool