Pincushion for Codex.

Visual feedback that lands in your IDE. Stakeholders drop pins on your live app; your Codex agent reads them via MCP — selector, DOM snippet, screenshot, viewport, thread, project context, and acceptance criteria — and ships the change from inside your editor.

The flow

1. Stakeholder drops a pin on any element of your live app via the Chrome extension
2. You open Codex, run /pins or /implement
3. Codex agent receives the pin's full work packet via MCP — no copy-pasting screenshots, no translating feedback into prompts
4. Agent finds the source code, makes the change, calls fix_and_resolve with branch + PR URL
5. Deploy hook auto-resolves the pin on production. Stakeholder sees "Resolved in PR #142, deployed at 3:24pm"

Connect Codex (60 seconds)

In your repo, run the guided setup and choose Codex explicitly, or add the MCP server through the Codex CLI:

codex mcp add pincushion -- npx pincushion-mcp --project-dir .

Restart Codex, then ask the agent to call get_actionable_pins. The command registers configuration; it does not prove the agent is connected. Review an existing AI or human finding, approve that exact pin, then ask the agent to implement it.

If this setup came from a Crit report, run npx pincushion-mcp claim <report-token>; it resumes the report’s project only after authenticated access is confirmed. The Chrome extension is optional and is only needed to place new browser pins.

Slash commands available after install

• /pins — show all open pins on this project
• /implement — run the full approved-pin loop (claim → fix → resolve)
• /resolve <id> — claim and close one pin
• /critique <url> — Pincushion AI drops ≤3 UI/copy/a11y pins on the page
• /feedback-summary — project overview

Existing feedback becomes a concrete agent task with its selector, DOM, screenshot, viewport, and thread in one MCP call.