Every plugin for OpenAI Codex on GitHub — repos with a .codex-plugin/plugin.json or a Codex marketplace. Install with codex plugin marketplace add owner/repo then codex plugin add name@marketplace. Many work in both tools; see Claude Code plugins.
Research stocks and market developments across China, Japan, and the United States using available public information. Explore stock moves, macro context, market sentiment, and technical structure, and prepare pre-market briefs or market recaps. Organize findings with source references, distinguish confirmed facts from interpretation, and identify information gaps.
Build Android apps with workflows for Kotlin and Compose setup, Material 3 UI, refactors, App Actions, emulator browser previews, debugging, performance profiling, and leak investigation.
A disciplined control loop for multi-window agent work across a tmux-resident fleet of Claude Code sessions: a Controller window claims requirement packages, plans task packages, prepares deliveries, imports immutable target results, obtains independent Test validation and reviews the evidence before completing and archiving a Demand. Explicit on-disk state roots keep every step auditable; Agents perform the host actions, Wakeflow owns the content and the verification. Supports macOS only (local file system); Linux and Windows are not supported.
Kabo client plugin: skill search/download/verification, skill-runner execution orchestration, dynamic meta-guidance (injected after Ed25519 signature verification). Usage telemetry is limited to event-level metadata (which skill ran and whether it succeeded); no prompt, tool or skill-output content is collected.
ARC agent plugin for Claude Code, Codex CLI, and Grok Build: AFS, blocklets, local MCP (stdio), and idempotent CLI bootstrap (ensure-arc.sh). Install from ArcBlock/agent-skills; v1 is local-only — remote MCP is #6432.
Refactoring review through 7 sourced refactoring personas (Fowler, Uncle Bob, Feathers, Beck, Metz, Ousterhout, Tornhill): scans smells + git hotspots, synthesizes a safety-first plan, then an adversary red-teams it.
Weave is a dependency-free engineering layer for Claude Code and Codex. It wraps shell commands to strip noisy successful output while always preserving failures in full, keeps a compact .weave/state.md handoff so context survives across sessions and between agents, and applies one consistent minimalism-and-verification policy so agents reuse existing code, prefer native features over new dependencies, and back every completion claim with real evidence.