---
title: "evofence"
description: "Inspect EvoFence's audit ledger and invoke evidence-gated evolution runs."
canonical: https://agentpluginsdirectory.com/plugins/evofence
last-updated: 2026-10-23
---

# evofence
Inspect EvoFence's audit ledger and invoke evidence-gated evolution runs.
- Slug: evofence
- Publisher: Calvin Xia
- Repository: https://github.com/Calvin-Xia/EvoFence
- Manifest: integrations/codex/plugin.json
- Version: 0.5.1
- License: MIT
- Category (editorial): other
- Skills: 2 (inspect-ledger, run-evolution)
- MCP servers: 0
- Stars: 1
- Repository created: 2026-09-24
- Repository last pushed: 2026-10-10
- Publisher type: User
- Listing: https://agentpluginsdirectory.com/plugins/evofence
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## What evofence does, in the publisher's words

This portable Codex plugin provides the namespaced skills $evofence:inspect-ledger and $evofence:run-evolution. Codex prefixes plugin skills with the plugin name, so the descriptive command names remain scoped to EvoFence in the skill list. It works with Codex CLI and Codex in the ChatGPT desktop app through a marketplace. The plugin contains instructions only; EvoFence's CLI remains the control plane for contracts, evidence, budgets, and candidate acceptance.

The Codex adapter does not support budgets.max_usd. Preflight refuses it with UNSUPPORTED_COST_BUDGET: Codex does not provide complete, verifiable USD telemetry. Set budgets.max_usd to null or use the Claude Code or Pi adapter. The alternative is to set budgets.max_usd to null, or choose Claude Code or Pi; do not bypass the refusal.

Restart Codex after installation if the skills do not appear.

From the project README, punctuation lightly normalized. Full text: https://raw.githubusercontent.com/Calvin-Xia/EvoFence/HEAD/integrations/codex/README.md