Productivity plugins

Reconstruct legacy systems from code and maintain source-backed architecture understanding for implementation and debugging.

1 skill

Guide coding agents to locate code with compact jgrep searches.

1 skill

Turn AI conversations into illustrated learning notes and save them to ZNote.

1 skill

Physical status bar for Claude Code and Codex tasks on an ESP32-S3 with an AMOLED screen

hooks

Compress videos to HEVC or H.264 with ffmpeg or macOS avconvert, skip files that are already efficient, verify every result, and never touch the original.

1 skill

Install and operate Auto (Jev), designed for Codex GPT-6 Luna, GPT-6.1 Sol, and GPT-6 Astra.

1 skill

Narrated walkthrough videos of a running web app: ElevenLabs voiceover, Playwright captures paced to the voice, HTML cards, ffmpeg assembly.

1 skill

Persistent project agents, temporary helpers and a plain-language HTML progress view.

1 skill

Quotes your own past corrections back to the agent before it answers, reports how often you had to correct it, and makes the orchestrator name a model for every subagent. Claude Code and Codex.

hooks

AI photobook maker: turns phone photos, WhatsApp images, HEIC and scanned prints into a story-driven, print-ready photo book PDF (chapters, consistent colour grading, message pages, print-quality check) and guides or drives the order at Print&Bind (printenbind.nl) or any PDF printer.

1 skill

A second model narrates what the agent does, reviews it, and lets you correct it mid-turn.

1 skill

Drive UV Studio+ in Chrome for a Quansheng UV-K5 V3 / UV-K1 on F4HWN firmware: navigate, configure memories and settings, and hardware-test builds.

1 skill

AI fact checker for Claude: checks documents claim by claim, recomputes figures and verifies quotes by script.

1 skill

Open source (MIT): one tiny status table and one clickable next step when you say STATUS.

1 skill

AgentZero plugins for Codex and DeepSeek Harness (DSH): register agents, collaborate on tasks, manage assets and explore the shared town.

1 skill

Generate native .drawio diagrams from Codex using Mermaid or XML, with optional ELK layout, PNG/SVG/PDF export, and browser URL output.

1 skill

One skill, 15 tools, sparse-by-design config. Direct OOXML mutation for Word documents — restyle by pattern, install styles and numbering, insert content with captions and cross-references, audit conformance.

ECC with native Codex hooks

Extract and display LinkedIn post statistics including impressions, reactions, comments, and reposts. Connects to Arc browser via CDP to read recent activity.

Clean and improve podcast transcriptions by removing filler words, fixing grammar errors, and applying ASR corrections. Supports both automated cleaning and manual reasoning-based review workflows for TXT and SRT subtitle formats.

Universal AI-powered task and backlog management with narrative continuity: handovers, issues, areas, finite epics (done_when). A SQLite store (.taskmaster/local/store.db) is the runtime authority; backlog.yaml and the per-entity markdown files are a Git-facing projection, so concurrent agents no longer overwrite each other's writes. New projects start on the native store (7.4; backlog_init(legacy_store=True) opts out), and existing checkouts cut over explicitly: one coordinator owns writes, reads take milliseconds, hand edits import via backlog_sync, and .taskmaster/ is committed with the managed git command. Existing v2/v3/v4 backlogs are adopted on first open (first open of a large backlog takes minutes). Kanban viewer with inline edit-in-UI served at the root URL for every project, in dark and light themes (7.2 re-skin: keyboard and screen-reader reachable, phone layouts). Terminal mods (7.3): a Taskmaster band above the prompt, a sidebar and tab row over review, handovers, session and restore panes. Skills and tool text rewritten for the Claude 5.5 models (7.4). Lessons subsystem removed in 4.0 (durable knowledge lives in assistant memory + CLAUDE.md/AGENTS.md). Requires: uv (https://docs.astral.sh/uv/)

Low-cognitive-load output styles for people with brain fog, attention difficulty, or fatigue.

Praxys training dashboard — MCP tools for training data, daily briefs, race forecasts, and AI plan management

AI Config Generator

Skills that guide AI coding agents to build, debug, and improve evals.

No description.

Local-first desktop control plane for governed coding-agent Outcomes.

Curated portable agents (curated-agents.txt): critic, reviewer, debugger, and specialist roles.

CI-native multi-agent review pack: risk-tiered reviewer fan-out plus a coordinator judge pass emitting one severity-rated review (critical-only blocking).

Claude Code edition: agent effort, mods and eval suites on top of agent-harness-kit