Agent Deck Comparison — AI skill for Claude Code

Agent-manager comparison

A public, sortable comparison table of AI coding-agent managers: tools that supervise a fleet of coding-agent sessions (Claude Code, Codex, and similar) across usage tracking, task sourcing, session organization, and remote/mobile access.


What this is

A single static HTML page (`index.html`) that loads `data/tools.json` and renders it as:

• Feature matrix: tools as columns, features as rows, matching the source comparison's cross table. Orca is the leftmost column. Click a tool's column header to sort feature rows by that tool's mark; click "Feature" to reset. A category filter, a feature search box, and a toggle to hide rows where every visible tool is "unknown."
• By tool: one row per tool (name, category, where agents run, license, checked-on date, source note, count of "yes" marks), sortable by any column.

Marks are rendered as small colored badges with text labels, not color alone, and the page works down to a 390px-wide phone screen with light/dark mode following the OS setting.

Snapshot and verification caveat

**Last updated 2026-10-15.** Every cell comes from each tool's own documentation, website, or public posts at the date noted for that tool. It has not been independently re-verified beyond that single pass. Where a tool's own materials don't state something, the cell reads "unknown" rather than a guess.

**Pantheon is the author's own project** ([open source](https://github.com/dtiger1889-ops/pantheon), pre-release, published as-is) and is held to exactly the same rules as every other tool here, including its weak spots (for example: 2 agent harnesses so far, no native mobile app, a terminal-multiplexer stability issue still open as of the snapshot date).

How to view it

Open `index.html` directly in a browser, or serve the folder with any static file server, e.g.:

python -m http.server 8080

then visit `http://localhost:8