Writing & Editing

Claude Code skills, one directory each. Starts with humanize: rewrites prose so it reads as written by a person, and explains why each tell reads as machine-made.

Enterprise gateway for git-distributed AI agent skill marketplaces (Claude Code, Copilot, Cursor)

name: yotron-ai

Cross-agent skill for 29-source job-market census, audited dashboards, and personalized realistic role research

Verified, installable workflows powered by Scrollport tools.

Claude Code plugin that teaches the assistant the real @theokit/sdk API surface — Agent.create, Tool.create with Zod, streaming SDKMessage events, MCP servers, subagents, cron and the error hierarchy — so it writes correct SDK code instead of plausible code. npx @theokit/skill. Apache-2.0.

A deterministic Windows field guide for coding agents — evidence before action.

Maz Works Knowledge Vault — shared project memory, research, skills, and operating context for Manazir Hussain's work.

孙本位 —— 先把一段经历审计成一张账本，再写成孙体。不许编数字。账本密度跑分 + 结账单卡片 + 孙本位汇率表。

Official Claude, Codex, MCP, and agent skills for AgentMailer

Small, composable Claude Code plugins: anvil (a one-command interview-plan-execute-review pipeline), an adversarial five-advisor council, six vendored engineering skills, and a game save-file reverse-engineering method.

Ideaspace Protocol extension for Pi — durable awareness, deliberate capture, optional sync.

Financial Disclosure Analysis agent skill: extract traceable financial data from official disclosures and generate editable PPTX reports with PPT Master.

Turn sources into social-ready copy and self-contained visual briefs.

A Chinese AI Agent writing Skill powered by 12 social-psychology triggers — turn raw material into restrained, memorable and shareable narratives.

Evidence-grounded resume-to-job fit analysis using only the supplied resume and job description.

Jess's agent skills/plugins/etc

AI武汉话方言技能 | 武汉话字典 —— 让任何 AI/Agent 读一遍就会说地道武汉话，无需微调。含语音规律、高频词表、语气词用法、歇后语，以及外地人和大模型最容易犯的四个错。用 I-Lang 协议编写。Wuhan dialect skill for LLMs and AI agents: vocabulary, phonetics, sentence-final particles, density control. Drop-in prompt, no fine-tuning.

Gauntlet Mini: the gauntlet loop with a mini bill. Builders, blind critics, and a 50/75/95% usage-budget ladder so the loop converges instead of torching your limits

AI agent drift detection and recovery for Claude Code and OpenAI Codex—stop repeated tool-call, instruction-following, and output-format failures with permission-gated handovers.

Install once. Keep durable project handoffs, instant status, sealed versions, and multi-agent coordination across AI coding agents.

Pass a note from your AI to someone’s phone. No account.

The LinkedIn post skill for Claude that interviews you first. It cannot write anything you did not say.

察言观数 · Read public voices, track visible signals, and ground AI in traceable evidence.