Agent Plugins Marketplace
← All plugins

cept

v0.4.1

Proprioception for coding agents: slice recent Claude Code transcript, redact it, and ask for outside-in steering.

CodexClaude Code1 Skill1 MCP serverstdio

By Eidos AGILicense: MIT0 GitHub starsUpdated 2 weeks ago

Directory evidence

Runtimes
Codex and Claude Code
Parsed components
2 skill or MCP entries
Source updated
Oct 4, 2026
Manifest status
Canonical path parsed

The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →

Install cept for Codex and Claude Code

Installs for the current user
codex plugin marketplace add eidos-agi/eidos-marketplace
codex plugin marketplace upgrade eidos-agi
codex plugin add cept@eidos-agi

Paste and run these commands in a terminal with Codex. They add and refresh the eidos-agi catalog, then install this plugin.

Compatibility: the page URL and API slug “cept” remain stable.

  • Codex: cept@agent-plugin-marketplace → cept@eidos-agi

The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.

Get the source manually
git clone https://github.com/eidos-agi/eidos-marketplace

Clone the source repository, then follow its setup instructions to add the plugin to a compatible client. The plugin root is plugins/cept/.

Plugin files

plugins/cept/
├── .codex-plugin/plugin.json
├── .claude-plugin/plugin.json
├── skills/use-cept/SKILL.md
└── .mcp.json

Included Skills1

use-ceptskills/use-cept/SKILL.md

Use when the user asks to use Cept or cept, asks for outside-in steering, wants an agent session reviewed, is stuck or looping, needs debugging triage, or wants architecture/research guidance from the recent agent trajectory. This skill tells Codex to call the local Cept CLI or MCP tool first and treat Cept output as evidence to reconcile.

MCP servers1

ceptstdio
command
uv
args
run --directory ${CLAUDE_PLUGIN_ROOT}/. cept
env.CEPT_EMIT
hud,file:~/.cept/status.jsonl

MCP configuration uses runtime-provided plugin path placeholders such as ${PLUGIN_ROOT} or ${CLAUDE_PLUGIN_ROOT}. Review the manifest for the runtime-specific expansion rules.

Plugin manifests2

plugins/cept/.codex-plugin/plugin.json
{
  "name": "cept",
  "version": "0.4.1",
  "description": "Proprioception for coding agents: slice recent Claude Code transcript, redact it, and ask for outside-in steering.",
  "author": {
    "name": "Eidos AGI",
    "email": "[email protected]",
    "url": "https://github.com/eidos-agi"
  },
  "homepage": "https://github.com/eidos-agi/cept",
  "repository": "https://github.com/eidos-agi/cept",
  "license": "MIT",
  "keywords": [
    "mcp",
    "claude-code",
    "codex",
    "agent",
    "proprioception",
    "openrouter",
    "steering"
  ],
  "skills": "./skills/",
  "mcpServers": "./.mcp.json",
  "interface": {
    "displayName": "Cept",
    "shortDescription": "Give an AI proprioception over its recent work and next move.",
    "longDescription": "Cept gives Codex and Claude Code a deliberate pause point when the agent is stuck, looping, debugging, researching, or making an architectural choice. It locates the recent agent transcript, distills the last few minutes of work, attaches repo state, redacts secrets, and asks an OpenRouter-backed Perplexity model for structured guidance.",
    "developerName": "Eidos AGI",
    "category": "Developer Tools",
    "capabilities": [
      "CLI",
      "MCP",
      "Claude Code Session Review",
      "Debug Steering",
      "Research Steering",
      "Architecture Review",
      "Secret Redaction"
    ],
    "websiteURL": "https://github.com/eidos-agi/cept",
    "privacyPolicyURL": "https://github.com/eidos-agi/cept/blob/main/SECURITY.md",
    "termsOfServiceURL": "https://github.com/eidos-agi/cept",
    "defaultPrompt": [
      "Use Cept to review this stuck debugging loop.",
      "Ask Cept for outside-in architecture guidance.",
      "Dry-run Cept before sending a transcript packet."
    ],
    "brandColor": "#2563EB"
  }
}
plugins/cept/.claude-plugin/plugin.json
{
  "name": "cept",
  "version": "0.4.1",
  "description": "Proprioception for coding agents: slice recent Claude Code transcript, redact it, and ask for outside-in steering.",
  "author": {
    "name": "Eidos AGI",
    "url": "https://github.com/eidos-agi"
  },
  "homepage": "https://github.com/eidos-agi/cept",
  "repository": "https://github.com/eidos-agi/cept",
  "license": "MIT",
  "category": "agent-tools",
  "keywords": [
    "mcp",
    "claude-code",
    "agent",
    "proprioception",
    "openrouter",
    "steering"
  ],
  "skills": "./skills/",
  "interface": {
    "displayName": "Cept",
    "shortDescription": "Give an AI proprioception over its recent work and next move."
  },
  "mcpServers": {
    "cept": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "${CLAUDE_PLUGIN_ROOT}",
        "cept"
      ],
      "env": {
        "CEPT_EMIT": "hud,file:~/.cept/status.jsonl"
      }
    }
  }
}

For maintainers

If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.

[cept on Agent Plugins Marketplace](https://pluginsmp.com/plugins/cept)