coldHat
io.github.vince-gonzalezv0.1.0Updated Oct 11, 2026
Put on anyone's AI rules by GitHub username: fetches their hat.md and its hat code.
Overview
Lets an assistant fetch a person's public hat.md rules from GitHub by username and work under those rules for the chat.
- What it does
- coldHat serves hats: a hat.md file kept in a public GitHub repository named coldhat, owned by one person. The MCP connector exposes a tool named wear_hat and a prompt named open_sesame; the assistant reads the hat, answers with its hat code to prove it read it, and then follows those rules for the rest of the chat. Extra hats live at hats/ .md and are addressed as /u/ / . Dated facts under a Status heading are checked against today's date, and the assistant asks about any that have gone stale.
- When to use it
- Use it when you want a chat assistant to adopt a specific person's working rules, voice, or review style without pasting them in by hand. It suits chats that can reach the connector, and hats can also be copied and pasted into chats that cannot browse.
- Requirements
- A remote MCP endpoint; no package, runtime, account, key, or header is declared. The hat owner needs a public GitHub repository named coldhat containing hat.md. The separate Python toolchain for authoring, locking, and evaluating hats is installed with pip and uses ANTHROPIC_API_KEY or OPENAI_API_KEY for model runs.
Installation
In SourceWeft
- Open coldHat in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.
Other MCP clients
Add this to your client's mcpServers config.
{
"mcpServers": {
"coldhat": {
"type": "http",
"url": "https://coldhat.f-keys.com/mcp"
}
}
}README
coldHat puts your working rules on any AI chat in one line. Type this into Claude, ChatGPT or Grok:
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.
[Claude and Grok each put on the frontier hat from one typed line and answer with its hat code]
Recorded 2026-10-11 in Claude (incognito) and Grok (private). Both quote 33963b16, the frontier
hat's live code, which they can only know by reading the page. Models keep their own copy of a
page for a while, so after an edit the code they quote can trail the live one. A playful
version, with the main hat's joke opener, is in docs/demo.gif.
Try one now. Starter hats anyone can wear:
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 at /u/<user>/<name>. Dated facts under
## Status (- label: value · as of YYYY-MM-DD · check after N days) are checked by the
server against today's date, and the assistant asks about any that have gone stale.
Which chat boxes work was measured, not assumed.
Edit your hat in your own editor
publish checks the file, commits only that file, pushes, and waits until the live hat code
matches your local one.
Under the hood
worker/: the Cloudflare Worker behind coldhat.f-keys.com: hat pages, the editor,llms.txt, the MCP server. Edge-cached, rate-limited, every name checked before GitHub is asked. 37 tests run in workerd.src/coldhat/: the Python toolchain for authoring hats as cards, locking them, and measuring them with evals across models.
A hat (as cards)
A card:
load: always cards sit in the instructions. load: retrieve cards are fetched
when a task calls for them, so the hat can grow without filling the context.
Activation. open sesame puts on head, body and footer. open sesame head
puts on the voice alone. hat off removes it.
Integrity. coldhat lock records a sha256 for every file. A hat whose files
differ from the lock refuses to load. The fingerprint is a hash over all of
them, and every eval run records it, so a score always names the exact hat
that earned it.
Authority. A hat describes how to work. It grants no tools, permissions or access; the host's own rules and the user's live instructions outrank it.
Delivery modes
Retrieval is BM25 over card titles, tags and bodies; standard library only and deterministic.
Commands
run and grade exit 0 when every case passed, 1 when any failed, and 2 when
none failed but some could not be fully graded.
Credentials come from each SDK's own environment variables
(ANTHROPIC_API_KEY, OPENAI_API_KEY). Name the OpenAI model with --model
or COLDHAT_OPENAI_MODEL. Claude defaults to claude-opus-5-5, with the API's
refusal fallback enabled.
Evals
A suite is a JSON list of cases. Each case is a prompt, usually a bait question built to tempt the model into breaking one rule, and a list of checks:
voice_bans run on every reply: announced honesty, filler words, a closing
"want me to...?" menu.
A judge check with no judge model is not_run. A judge that errors is error.
Either one makes the case incomplete, which is never counted as a pass. An
empty reply fails outright, since it would otherwise pass every "must not
contain" check.
Tuning
- Run the suite on a model.
- Read the failures. Each one points at a card, the head, or a missing card.
- Edit,
coldhat lock, run again. coldhat report OLD NEWshows which cases flipped and confirms the hat changed.
Once a hat scores well, coldhat export turns the head's example answers, the
suite's reference answers and the replies from judged, passing runs into
training data. It only takes replies from runs of the current fingerprint, so
behavior from an older hat cannot leak in.
Chat-window run
coldhat build hats/vinceand copydist/vince/MASTER.mdinto the chat.coldhat sheet evals/vince.json --out claude.json, send each question in that chat, paste each reply into the file.- Repeat in a second model's chat with
--out gpt.json. coldhat grade hats/vince evals/vince.json claude.json --label claude, same forgpt.json.coldhat reporton both runs, andcoldhat agreeto see whether they took the same positions.
Security
A hat is instructions, not a credential, and open sesame is not a password. Commits pass a secret and private-term scrubber before they exist, CI runs it again with gitleaks over the full history, hats refuse to read outside their own folder, and the only tool a hat exposes is a read-only search over its own cards. Details and limits are in SECURITY.md.
To install the commit hook in a clone:
python tools/scrub.py --hash WORD prints the line to add for a term.
Tests
No network and no keys. Each gate is tested on input built to pass it and input built to fail it, and both provider adapters are driven through their tool loops with fake clients.
Part of F-Keys — independent hardware, software and internet products. See the working log and live status.
Source: README.md at commit 78c3bfd
Tools
0Version history
1- v0.1.0LatestOct 11, 2026
More servers like coldHat
LoopSkill
io.github.wisechef-ai
Lets an assistant search, install and compose AI agent skills, loops and bundles from the LoopSkill marketplace over MCP.
ToGO
io.github.togo-framework
Lets an assistant look up ToGO framework documentation, browse its plugin marketplace, and draft a plugin submission issue.
Laravilt
io.github.laravilt
A public, read-only MCP server that gives assistants the latest Laravilt documentation, package versions, changelogs and plugin guides.