惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

大猫的无限游戏
大猫的无限游戏
D
DataBreaches.Net
M
MIT News - Artificial intelligence
量子位
N
Netflix TechBlog - Medium
The Cloudflare Blog
The GitHub Blog
The GitHub Blog
P
Proofpoint News Feed
人人都是产品经理
人人都是产品经理
B
Blog RSS Feed
B
Blog
博客园_首页
博客园 - Franky
MyScale Blog
MyScale Blog
有赞技术团队
有赞技术团队
Apple Machine Learning Research
Apple Machine Learning Research
MongoDB | Blog
MongoDB | Blog
云风的 BLOG
云风的 BLOG
爱范儿
爱范儿
H
Help Net Security
Y
Y Combinator Blog
Stack Overflow Blog
Stack Overflow Blog
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
酷 壳 – CoolShell
酷 壳 – CoolShell

Hacker News: Show HN

PurrrrrFocus: Pomodoro Timer App - App Store Workflow Engine — Multi-Step Orchestration for Bun RapidPhoto: Pro Photo Editor App - App Store GitHub - DheerG/swarms: Achieve extraordinary results with claude code across a variety of tasks SPICE simulation → oscilloscope → verification with Claude Code — Lucas Gerads Show HN: VCoding – A 5 MB native Windows IDE with no dynamic dependencies Show HN: LLMs don't hallucinate because they're bad at math, it's the format GitHub - Agent-FM/agentfm-core: AgentFM is a peer-to-peer network that turns everyday computers into a decentralized AI supercomputer. AgentFM lets you run massive AI workloads directly across a global mesh of idle CPUs and GPUs. Show HN: Tracking Top US Science Olympiad Alumni over Last 25 Years GitHub - Potarix/agent-hub: One place to talk to all your agents Show HN: Runtime security for AI agents(injection,tool abuse, data exfiltration) GitHub - dubeyKartikay/lazyspotify: Terminal Spotify client for macOS and Linux GitHub - the-banana-tool/king-louie: Easy to use GUI Personal AI Assistant. Win/Linux/Mac. Show HN I made my vacation rental bookable by AI agents–no Airbnb, 0% commission GitHub - basteez/jsf-autoreload: maven plugin to enable hot reload on jsf projects uvm32/hosts/host-gdbstub at main · ringtailsoftware/uvm32 GitHub - labsai/EDDI: Config-driven engine that turns JSON into production-grade AI agents. Multi-agent orchestration, 12+ LLM providers, MCP/A2A protocols, RAG, persistent memory, and enterprise compliance (EU AI Act, GDPR, HIPAA). Built on Quarkus. GitHub - glitchnsec/fortyone-oss: AI Executive Assistant Platform Quickstart | Alien GitHub - muxshed/shed: One stream in, or many. Every destination, simultaneously. No cloud middleman, no per-channel fees, no limits. GitHub - ocrbase-hq/ocrbase: 📄 PDF/IMG ->.MD/JSON Document OCR API for PaddleOCR and GLMOCR. Self-hostable. GitHub - impactjo/home-memory: MCP server that lets your AI assistant remember everything about your home. GitHub - Sets88/dbcls: DbCls is a powerful terminal database client that supports various databases GitHub - neptun2000/heor-agent-mcp GitHub - SeanFDZ/macmind: Single-layer transformer in HyperTalk for the classic Macintosh RollQuation: Math Puzzles - Apps on Google Play GitHub - dropbox/witchcraft Show HN: Agent-cache – Multi-tier LLM/tool/session caching for Valkey and Redis GitHub - opentalon/opentalon: OpenTalon is an open-source platform built from the ground up in Go as a robust alternative to OpenClaw LinkedIn™ 职位抓取工具 - Chrome 应用商店
Implementing patio11's "Dangerous Professional" as a Clau...
tawb · 2026-04-28 · via Hacker News: Show HN

Patrick McKenzie (patio11) has written for years about a register he calls the Dangerous Professional. It came out of his consulting work with banks and fintech, where he sat on both sides of the glass: institutions trained to deflect, and customers trying to get money back, fix billing errors, dispute charges, recover deposits. The voice is the one you use when you write to a bank, a regulator, a company that owes you something. Calm. Formal. Cites the policy by section. Names the deadline. Never threatens. The email reads like it could come from someone with a lawyer, without ever saying so.

X avatar for @patio11

Patrick McKenzie@patio11

I’d grade this an A. (Memetically, being a Dangerous Professional means communicating in what might be a slightly adversarial context in a way which suggests that a bureaucracy take one’s concerns seriously and escalate them to someone empowered to resolve them swiftly.)

X avatar for @FiloSottile

Filippo Valsorda @filippo.abyssdomain.expert @FiloSottile

Hey @patio11, how's my Dangerous Professional interpretation?

3:08 AM · Aug 17, 2019

2 Replies · 28 Reposts · 271 Likes

Patio11’s framing: institutions sit inside a tower of assumed competence, and the people inside respond very differently to messages that sound like they came from inside the tower. He’s used it himself — and helped friends use it — to reverse fraudulent charges, recover wrongly withheld funds, fix medical billing errors, push back on denied insurance claims. The skill is partly the script — what to cite, what to ask for, when to escalate — and partly the register itself, which signals you’re not going to go away. Most people never learn any of this. Some pick it up from family or work; most learn it the hard way, after they’ve already lost the dispute by sounding too angry or too hopeful.

I’ve been developing a sort of meta-plugin: a framework for automating the creation of new agent plugins. It works as a consultant. It interviews me about the problem, makes me point at references for what good looks like, refuses to autonomously hunt for examples (it’d just codify whatever bad patterns it finds), and routes the work to the right extension point. The meta-plugin is built from several deep research dumps on best practices for plugin writing, plus my own curated experience and that of power users in my network.

For this plugin specifically, I ran automated research dumps from a curated list of links where patio11 has talked about Dangerous Professional, and fed that into the meta-plugin. The meta-plugin compiled it into a new skill purpose-built for dangerous-professional. I handed that skill to Claude in a single plan session and let it build the thing.

This was a trial for automated skill development. It went far enough that I want to develop this into a regular framework to latch onto.

I’ve been using it on:

  • Contractors working on my house

  • Banks and SaaS support teams fighting overage charges, billing issues and bugs

  • Most impactfully, on my home insurance provider after a historic hail storm destroyed roofs across the metro area, including ours.

  1. Context dump the situation and all the supporting data into the agent. State what you actually want as an ideal outcome.

  2. Then ask the agent for feedback on the goal itself. Most of the time you’re wrong about what you’re asking for, or your aims are more emotional than they need to be.

  3. Workshop a few framings for the real goal. Pick one.

  4. Have it write the email.

Patio11 on the Dangerous Professional

Podcasts

  • Complex Systems — patio11’s own show. DP-adjacent material runs through it.

The plugin

No posts