#235235th of 455 in Skills, plugins and prompts
agent-purge
Local CLI and web console for managing one persona prompt across coding agents
- Stars
- 1.7k
- License
- MIT
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- JavaScript
107 on 20 Oct 20261,702 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
agent-purge is a Node.js tool that keeps persona prompts in Markdown files under ~/.agent-purge/ and applies them to coding agents such as Codex, Claude Code, OpenCode, pi, Kilo Code and Antigravity. It offers four modes (replace, strict, append, off) and several channels: one-off launch, managed config install, PATH shim, slash extension and per-project rules. A local web console and a CLI share the same state, and managed writes are recorded so they can be reverted.
Who it is for: Developers using several coding agents who want one shared persona prompt
Strengths
- Per-agent persona or mode override, with a global default to fall back on
- Records managed writes; pause, disarm and clean revert them, and conflicts raise errors
- Compatibility table lists tested host versions and whether same-session hot switching was verified
- Persona files live outside the program directory, so upgrades do not overwrite them
Weaknesses
- Hot switching is unverified for Codex, Claude Code, CodeBuddy and Grok; they need a new session
- macOS is not yet verified on real hardware; regression runs cover Windows and WSL/Linux
- The same persona sent through several channels may be duplicated
- Persona text is sent to the model provider with each host request
What it needs
- no GPU
- Needs Node.js 20+, a supported coding agent (Codex, Claude Code, OpenCode, pi, etc.)
- port 7788
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in Skills, plugins and prompts
See all 455| Rank | Project | Score |
|---|---|---|
| 1 | claude-memPersistent cross-session memory plugin for Claude Code and other coding agents | 97 out of 100 |
| 2 | impeccableDesign skill with 24 slash commands for AI coding agents' frontend output | 97 out of 100 |
| 3 | gstackMarkdown slash-command skills that give Claude Code planning, review, QA and ship roles | 96 out of 100 |
| #4 | rufloAgent harness adding swarms, memory and hooks to Claude Code and Codex | 96 out of 100 |
| #5 | cavemanShortens coding-agent replies and compresses tool output before the model reads it | 95 out of 100 |