Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/nweii/agent-stuff/things-appnpx skills add nweii/agent-stuff --skill things-appgit clone --depth 1 https://github.com/nweii/agent-stuffWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00110 | $0.03258 |
| Opus 5 | $0.00055 | $0.01629 |
| Sonnet 5 | $0.00022 | $0.00652 |
| Haiku 4.5 | $0.00011 | $0.00326 |
Grade A, and why
things-app scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 142 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Things 3
Drive Things 3 (tasks, projects, areas) through whichever modality fits the environment. Prefer the things CLI whenever there's command-line access to Things — it reads and writes the local database directly, so it's fast and needs no network round-trip. Where there's no CLI access, the fallbacks are: a connected Things MCP (full read + write that syncs across devices, from anywhere), the things:/// URL scheme (any device with Things, write-only), and email-to-Things (unattended, Inbox-only capture).
A Things MCP — a self-hosted server, or an official one Cultured Code may ship — plays the role the CLI plays locally, but reachable from anywhere, so it's the first fallback when there's no command-line access. Match its tools by their things_* base names; the MCP host prepends its own namespace, so the full identifier varies by client.
Pick a modality first
| Modality | Read state? | Schedule & update? | Unattended? | From phone? |
|---|---|---|---|---|
CLI (things binary) |
Yes | Yes | On the Mac only | No |
URL scheme (things:///) |
No | Yes (update/json need auth token) |
No — someone must open the URL | Yes |
| Email-to-Things | No | No (Inbox capture only) | Yes | Yes (send an email) |
The real axis is attended vs unattended, not headless. A headless session someone is watching can still use the URL scheme — the agent prints the link, the user opens it. Only a routine with no one to act on its output is limited to email.
Decision rules:
- Command-line access to Things (a terminal with the local Things database) → use the CLI. The default: reads (Inbox, Today, search, projects, areas, tags) and the full write surface, all against the local DB.
- No CLI access, but a Things MCP is connected → use it. Some setups expose Things through an MCP server (self-hosted, or a future official one); its
things_*tools give full read + write that syncs across devices, with no Mac or device interaction required. The path for headless, remote, or cloud-routine environments. - No terminal, but the user can act on the output → build a URL. You construct the
things:///...link; the user (or an iOS Shortcut) opens it on a device with Things to execute. This holds even when the agent itself is headless — being headless doesn't rule out the URL scheme; only the absence of anyone to open the link does. The mobile and attended-remote write/schedule path. - Fully unattended (a remote/NAS routine with no one watching) → email-to-Things. The only path that needs neither a Mac nor a human. Lands in the Inbox only — no scheduling, no project placement, no updates. Good for dropping raw captures that get triaged later on desktop.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 142 lines · 110 tokens per session scan A ab3467339af9
things-app is a skill published in the GitHub repository nweii/agent-stuff (8 stars, last pushed 14d ago), licensed MIT. It adds 110 tokens to every session and 3,258 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
hr-onboarding
A new-hire onboarding plan as a single page — first week schedule, buddy + manager intro, learning track, equipment checklist, and "you're set when…" outcomes. Use when the brief mentions "onboarding", "new hire", "first week plan", or "入职".
lark-im
飞书即时通讯：收发消息和管理群聊。发送和回复消息、搜索聊天记录、管理群聊成员、上传下载图片和文件、管理表情回复、发送应用内/短信/电话加急、发送和处理交互卡片（Interactive Card）、监听卡片按钮回调（card.action.trigger）。当用户需要发消息、查看或搜索聊天记录、下载聊天中的文件、查看群成员、搜索群、创建群聊或话题群、管理标记数据、管理 Feed 置顶（添加/移除/查询置顶会话）、管理标签数据、处理卡片回调时使用。.
feishu
Work with Feishu or Lark bots, docs, sheets, bitables, approval flows, and OpenAPI/MCP setup without hardcoding credentials.
reflect
Review recent work, find repeated workflow patterns, and suggest reusable skills, agents, commands, config changes, or playbooks. Use when the user asks to learn from past sessions, improve recurring workflows, or identify what should be turned into reusable agent instructions.
mochi-remind
Handle due reminders — notify the user with natural language and mark them done.
organize-threads
猫猫辅助整理未分类 thread，分析标题和元数据，建议合适的标签。 Use when: 用户说"帮我整理"、"分类 thread"、点击整理按钮。 Not for: 删除/编辑标签本身。 Output: 按 thread 的标签建议列表。.