Operate multiple coding agents from one Emacs workflow with AI Code Interface
Give Emacs users one workflow layer for Codex, Claude Code, Gemini CLI, Cursor CLI, opencode, and other coding-agent backends.
npx skills add agentskillexchange/skills --skill operate-multiple-coding-agents-from-one-emacs-workflow-with-ai-code-interface
Use AI Code Interface when an Emacs-based developer wants a stable agent-operator workflow while switching between coding-agent backends. From one transient menu, the operator can start or resume sessions, send scoped code-change prompts, attach files and screenshots, run guided TDD and refactoring flows, review diffs, manage Git context, and preserve session handoff notes. Invoke it when the user is actively doing AI-assisted coding from Emacs and needs repeatable session, context, review, and test workflows across backends instead of opening each product normally. The scope boundary is the Emacs workflow layer around already-installed coding agents; it does not replace Codex, Claude Code, Gemini CLI, Cursor CLI, opencode, or other runtimes, and it is not a general Emacs package listing.
What this skill actually does
Inputs and prerequisites: Emacs 29.1+, MELPA or straight.el, one supported coding-agent CLI, vterm/eat/ghostty or another configured terminal backend.
Setup notes: Enable MELPA, refresh package contents, install the ai-code package with M-x package-install, configure a backend such as (ai-code-set-backend ‘codex), and bind ai-code-menu, for example to C-c a.
Source and verification boundary: use https://github.com/tninja/ai-code-interface.el as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.