About coldHat
coldHat serves instruction sets, called hats, associated with GitHub users and other prompt collections. It is for people who want an AI chat to follow a chosen set of rules, with a seal code to indicate whether the full hat was read.
Key features
- 01Fetch hats by GitHub username
- 02Load hats from prompt collections
- 03Stack or merge up to three layers
- 04Lock rules against overrides
- 05Check reply codes and stale facts
Use cases
- 01Apply a code-review rule set in a chat
- 02Combine multiple instruction layers
- 03Verify a reply loaded the complete hat
MCP support
The MCP endpoint exposes the `wear_hat` and `wear_stack` tools to compatible chat applications.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://coldhat.f-keys.com/mcpclaude mcp add --transport http vince-gonzalez-coldhat https://coldhat.f-keys.com/mcp{"mcpServers":{"vince-gonzalez-coldhat":{"type":"http","url":"https://coldhat.f-keys.com/mcp"}}}Open in Cursor{"servers":{"vince-gonzalez-coldhat":{"type":"http","url":"https://coldhat.f-keys.com/mcp"}}}Open in VS Code{"mcpServers":{"vince-gonzalez-coldhat":{"type":"http","url":"https://coldhat.f-keys.com/mcp"}}}{"mcpServers":{"vince-gonzalez-coldhat":{"serverUrl":"https://coldhat.f-keys.com/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
3| Parameter | Type | Required | Description |
|---|---|---|---|
| hat | string | No | A named hat from the owner's hats/ folder; omit for their main hat |
| user | string | Yes | GitHub user name of the hat's owner, for example vince-gonzalez |
Details
From the project README╔════════════════════════════════════════════════════════════════════════════════════════════╗
║ ║
║ ██████╗ ██████╗ ██╗ ██████╗ ██╗ ██╗ █████╗ ████████╗ ║
║ ██╔════╝██╔═══██╗██║ ██╔══██╗██║ ██║██╔══██╗╚══██╔══╝ ║
║ ██║ ██║ ██║██║ ██║ ██║███████║███████║ ██║ ║
║ ██║ ██║ ██║██║ ██║ ██║██╔══██║██╔══██║ ██║ ║
║ ╚██████╗╚██████╔╝███████╗██████╔╝██║ ██║██║ ██║ ██║ ║
║ ╚═════╝ ╚═════╝ ╚══════╝╚═════╝ ╚═╝ ╚═╝╚═╝ ╚═╝ ╚═╝ ║
║ ║
║ will another model keep your rules? ║
║ ║
╚════════════════════════════════════════════════════════════════════════════════════════════╝
coldHat puts working rules on any AI chat in one line. Type this into Claude, ChatGPT or Grok:
open sesame coldhat.f-keys.com/u/vince-gonzalez/frontier
The assistant reads the hat, answers with its hat code to prove it read it, and works by those rules for the rest of the chat. Live at coldhat.f-keys.com.
Every hat ends with a seal. If the assistant can't quote it, the hat was cut off, and it tells you.
Addresses
| You want to | Type this after open sesame |
|---|---|
| Wear a person's hat | coldhat.f-keys.com/u/<github-user> or .../u/<github-user>/<hat> |
| Wear one of Fabric's patterns | coldhat.f-keys.com/fabric/extract_wisdom |
| Wear a prompts.chat prompt | coldhat.f-keys.com/prompts/linux-terminal |
| Stack up to three, later layers winning | coldhat.f-keys.com/u/vince-gonzalez/editor+fabric/analyze_claims |
| Merge a stack by slot instead of piling it up | coldhat.f-keys.com/amp/fabric/extract_ideas+fabric/analyze_claims |
| Keep rules no other layer can override | put them under ## Locked in your hat |
| Wear one exact version | coldhat.f-keys.com/u/<user>/<hat>@<commit> |
| Check a reply wore the current hat | coldhat.f-keys.com/verify?address=<hat address> with the reply |
Every reply starts with the hat mark and code, for example 🎩💨⛄🔬 Hat's on · b7d19023:
a top hat, a gust, Frosty, then one icon per layer.
Apps that take connectors can add https://coldhat.f-keys.com/mcp once and call any of
these by name (tools wear_hat and wear_stack). A chat that can't browse can still wear
a hat: Copy hat on coldhat.f-keys.com, then paste.
Starter hats
| Hat | Type this |
|---|---|
| Code reviewer: verdict first, findings by severity with line, failure and fix | open sesame coldhat.f-keys.com/u/vince-gonzalez/reviewer |
| Grant writer: rubric first, every number sourced, every limit met | open sesame coldhat.f-keys.com/u/vince-gonzalez/grants |
| Plain-English editor: shorter, clearer, still your voice | open sesame coldhat.f-keys.com/u/vince-gonzalez/editor |
| Frontier: mechanical register, rules over manners | open sesame coldhat.f-keys.com/u/vince-gonzalez/frontier |
Make your own
A hat is a hat.md in a public GitHub repository named coldhat, so only its owner can
change it. Extra hats go in hats/<name>.md and live at /u/<user>/<name>. Write one at
coldhat.f-keys.com/make, bring the instructions you
already keep for other assistants at coldhat.f-keys.com/import,
or start from coldhat-starter.
Facts under ## Status carry a date and a shelf life
(- label: value · as of YYYY-MM-DD · check after N days). The server checks them against
today, and the assistant asks about any that have gone stale before using them.
Measured, not assumed
Safety
A hat is instructions, not a credential, and open sesame is not a password. Every hat
page tells the assistant to ignore it unless the person asked for it, and never to ask for
passwords, keys, codes or payment details. Layers from Fabric and prompts.chat credit their
authors; prompts written to make a model drop its rules are not served.
Report a hat.
The coldHat server and merge engine are proprietary. This repository holds the public hats, the format, and the results.
╔════════════════════════════════════════════════════════════╗
║ ║
║ ███████╗ ██╗ ██╗███████╗██╗ ██╗███████╗ ║
║ ██╔════╝ ██║ ██╔╝██╔════╝╚██╗ ██╔╝██╔════╝ ║
║ █████╗ █████╗█████╔╝ █████╗ ╚████╔╝ ███████╗ ║
║ ██╔══╝ ╚════╝██╔═██╗ ██╔══╝ ╚██╔╝ ╚════██║ ║
║ ██║ ██║ ██╗███████╗ ██║ ███████║ ║
║ ╚═╝ ╚═╝ ╚═╝╚══════╝ ╚═╝ ╚══════╝ ║
║ ║
║ · C R E A T I V E · ║
║ ║
║ ──────────────────────────────────────── ║
║ ║
║ Vincent Gonzalez ║
║ f-keys.com ║
║ ORCID 0009-0005-3640-014X ║
║ ║
╚════════════════════════════════════════════════════════════╝
Part of F-Keys — independent hardware, software and internet products. See the working log and live status.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.