Ecosystem front door

Ecosystem map for Codex, Claude Code, and MCP.

CortexPilot is a command tower for Codex and Claude Code workflows with a real read-only MCP surface. This page exists to make that binding easy to explain, while keeping adjacent tools in the right comparison layer instead of inflating them into fake hero claims.

Need the sharper “how would my coding-agent team actually use it?” answer? Pair this page with the integration guide and the compatibility matrix, then the skills quickstart.

Primary: Codex workflows Primary: Claude Code workflows Protocol: read-only MCP Adjacent: OpenHands / OpenCode

Why these names stay in the front door

CortexPilot already ships real operator-facing surfaces for Codex-style and Claude Code-style workflows: Command Tower, Workflow Cases, Proof & Replay, explainable operator briefs, and structured read-only MCP inspection. That is why these names belong in the public discovery layer.

Primary workflow binding

Codex

Use CortexPilot when a Codex-led workflow needs one governed command tower, one durable case record, and one replayable proof path.

Primary workflow binding

Claude Code

The same operator surface works for Claude Code loops that need approvals, evidence, and a visible workflow trail instead of disconnected chat output.

Protocol surface

Read-only MCP

MCP is already real here, but the truthful public contract stays read-only. External tools can inspect workflow truth without taking control they do not yet have.

Ecosystem surfaces and how they fit into the CortexPilot product story.
Surface Where it fits Why it matters Current truthful boundary
Codex Hero and first-screen discovery High-intent searchers should immediately see that CortexPilot is built for governed coding workflows, not generic AI demos. Supported as a primary workflow audience.
Claude Code Hero and first-screen discovery Pairs with Codex so the front door speaks to the two primary workflow audiences without inventing partnerships. Supported as a primary workflow audience.
MCP Hero-adjacent protocol layer Signals a real integration boundary for inspection, proofs, approvals, and workflow reads. Read-only only; no write-capable MCP claims.
OpenHands Adjacent ecosystem / comparison layer Useful for “where does this fit in the broader agent stack?” discovery without turning it into a co-brand. Comparison-only; not elevated into the main hero.
OpenCode Adjacent ecosystem / comparison layer Helps transition-minded searchers understand overlap without muddying the primary story. Comparison-only; keep it out of the hero.
OpenClaw Trend / landscape only People may search around the same buzz cluster, but it points at a different product category. Not part of the front door.

Adjacent tools still matter, just not in the hero

Adjacent ecosystem

OpenHands

Keep it in the comparison layer when you want to explain how CortexPilot sits beside broader agent-stack orchestration and execution tooling.

Adjacent ecosystem

OpenCode

Use it in migration or comparison copy, not as a primary product anchor. The public front door should stay focused on Codex, Claude Code, and MCP.

Landscape only

OpenClaw

Keep it in trend/landscape copy only. It lives close enough to the keyword cloud to mention, but far enough from the core product story to keep out of the hero.

Where builder and API entrypoints fit

The ecosystem story does not stop at names. CortexPilot also ships thin builder entrypoints for API client helpers, contract-facing types, and a shared UI substrate. Those belong in the onboarding ladder, not in the hero.

Official platform reality

The ecosystem names in this front door are real, but their platform surfaces are not interchangeable. Keep workflow integrations, protocol surfaces, plugin systems, and publishing workflows separate instead of collapsing them into one vague “plugin” story.

Surface What is real today What we should not overclaim
Codex Codex has real workflow integration surfaces such as the Codex SDK, GitHub Action, and Slack integration. Do not turn those integration surfaces into a claim that CortexPilot is already an official Codex plugin or directory listing.
Claude Code Claude Code has official MCP, subagent, hook, and project-configuration surfaces for governed coding workflows. Do not turn those workflow surfaces into a claim that CortexPilot is already an official Claude Code plugin or marketplace listing.
MCP MCP is an open protocol, and there is now an official MCP Registry for servers. Do not collapse “MCP Registry” into “universal plugin store” language; registry/server discovery is not the same thing as a client marketplace.
OpenClaw OpenClaw has official plugins and skills surfaces, including a public skills registry, so plugin-ready wording is real in that ecosystem. Do not turn that into a claim that CortexPilot itself already ships as an official OpenClaw plugin unless a real package and tested path are published.

CortexPilot can truthfully map itself to these surfaces, but it should not pretend that repo-owned builder/docs assets are already a live plugin marketplace, a Codex/Claude official plugin, or a hosted platform product.

Current guardrails

If you want the product-surface version of this map, continue to the compatibility matrix, AI + MCP + API surfaces or the builder quickstart.