Back to browse
Codeman – a blunt launcher forcing you to pick a Codex permission level

Codeman – a blunt launcher forcing you to pick a Codex permission level

by shabers·Feb 13, 2026·1 point·0 comments

AI Analysis

●●SolidNiche GemSolve My Problem
The Take

It forces explicit, blunt choices — seven named modes with emoji risk markers and a TTY confirmation panel — which is exactly the kind of friction you want before giving an AI destructive power. Useful workflow features like resume-by-session-UUID, notify-test for Slack/Discord, and generated shell completions show attention to real usage. The tradeoff: it's narrowly targeted to Codex users and it’s not obvious whether it enforces stronger sandboxing or just makes you click the right flag.

Target Audience

Developers and engineers who run Codex/AI coding sessions and want an explicit, safer permission workflow (security-conscious devs, AI tool users)

Post Description

I kept getting burned by “what permissions am I running with?” so I built Codeman: a thin launcher around codex that forces an explicit security level each run (read-only, orkspace-write, networked, full), with a clear confirmation panel before higher-risk modes. It also supports resuming by session UUID and optional Slack/Discord webhook notifications. Repo: https://github.com/shabo/codeman Feedback welcome: naming, UX, and whether the level breakdown matches how you actually work with Codex.

Similar Projects

CodexL – All-in-one launcher for Codex

Polished Codex wrapper, but another launcher in a category that doesn't need one.

SlickNiche Gem
musistudio
207d ago
AI/ML●●Solid

Brute-force startup ideation with the Ralph Loop

The core trick is simple and effective: let an agent iterate questions against a defined domain overnight and surface hundreds of candidly-annotated ideas you can scan through later. It nails the “fire-and-forget” idea dump and domain steering (tell it to focus on agencies or cybersecurity and it pivots), but it’s still essentially a convenience wrapper around an existing agent pattern — useful for volume and pattern recognition, less convincing on long-term validation or downstream filtering.

Rabbit HoleNiche Gem
bothlabs
203mo ago