v1.0.0

Changelog

All notable changes to this project are documented here.

[1.0.0] — 2026-06-14

First public release. Designed and built end-to-end by an AI agent (Claude, in Claude Code).

Added

• Inventory of every agent element across Claude Code (~/.claude) and
  Codex (~/.codex): skills, subagents, slash commands, plugins, MCP servers, hooks.
• Provider switcher (All · Claude · Codex) and per-item provider badges.
• Sessions view — live / resumable / stale, with token usage, recall commands, Reveal
  in Finder, and guard-railed cleanup to the Trash.
• Insights — token & cost analytics across both agents (spend by project and model,
  Claude and GPT side by side) with an activity heatmap.
• Hooks / automation audit — hooks by event, Codex command rules, and ~/.claude
  sweep markers.
• Relationships — "who can use what" across subagents, plugins, and projects.
• Markdown previews for SKILL.md / agent / command / plan bodies, with a Rendered/Raw toggle.
• Manage plugins & skills — enable/disable from the UI, path-locked and backed up.
• Codex coverage — config-driven trust levels, command rules, best-effort live detection,
  and codex resume recall.
• Menu-bar extra, Welcome and Help/About sheets, and a native app icon.
• Distribution: ./build.sh release --dist produces a release zip.