A guide for running fold locally and using it through the Model Context Protocol, a standard way for clients to call tools and services. It starts fold with a test upstream so you can inspect gateway behavior from a client's perspective.

Use it to start the local development stack, check service health, compare fold with its upstream, and test namespaced responses.

It helps investigate bugs or surprising behavior that ordinary automated tests may not reveal.

How fold-run/fold configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to fold-run/fold. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/fold-run/fold/main/.claude/skills/dogfood/SKILL.md

Clone the repo

git clone --depth 1 https://github.com/fold-run/fold

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/fold-run/fold/dogfood/github.svg)](https://agentmods.dev/skills/fold-run/fold/dogfood)

Your own site

<a href="https://agentmods.dev/skills/fold-run/fold/dogfood"><img src="https://agentmods.dev/badge/skills/fold-run/fold/dogfood/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/fold-run/fold/dogfood"><img src="https://agentmods.dev/badge/skills/fold-run/fold/dogfood.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 69 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 847 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What 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 5.1	$0.00069	$0.00847
Opus 5.5	$0.00028	$0.00339
Sonnet 5.5	$0.00014	$0.00169
Haiku 4.5	$0.00007	$0.00085

Measured 2mo ago against content hash 5c07d798b2c4, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

dogfood 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 2mo 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.

.claude/skills/dogfood/SKILL.md · 70 lines

How it starts

The opening of the file, as written. The whole thing — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Running fold and using it

fold's central claim is about what a client experiences, and the fastest way to check a claim about experience is to have the experience.

make dev-up       # everything-server on :8098, fold on :8099
make dev-status   # the health body, upstream by upstream
make dev-down

scripts/dev-stack.sh builds fold from the working tree, so it serves the code you are editing. The upstream is the reference everything-server — the same peer make conformance fronts, so what you see here and what CI gates on are the same server. It is federated under the namespace demo rather than passthrough, deliberately: namespacing is the behavior most worth seeing from the outside (demo__echo, not echo).

Ports are overridable and are not the gateway's own default 8080, which is a popular squat — Docker Desktop takes it. The script refuses to start on a busy port and names the process holding it rather than failing obscurely.

Two clients

The Inspector — one method, scriptable, no session state to reason about. /mcp-inspector covers it, including the invisibility diff: the same method against fold and against the upstream directly, compared.

This Claude Code session — .mcp.json at the repo root points at http://127.0.0.1:8099/mcp, so with the stack up and the server approved, the session's own tool list contains fold's federated tools. That is a genuine end-to-end client, with all the awkwardness a real one has.

The gateway must already be running: fold serves streamable HTTP and has no downstream stdio mode, so nothing here is self-starting. If the server shows as failed, the stack is down — make dev-up, then reconnect.

What to look at once it is up

Namespacing round-trip. Tools list as demo__*; a call with the namespaced name reaches the upstream under its bare name. A call with the bare name answers -31043.
The console at http://127.0.0.1:8099/console/ — the federation view an operator gets, served from the vendored assets (/console-sync).
Health, which answers 503 with a body when an upstream is down. Stop the upstream and watch the breaker move; "listening" and "ready" are different claims and this is where the difference is visible.
The audit trail on stdout in $TMPDIR/fold-dev-stack/fold.log — one event per terminal response, denials included.

Read the full file on GitHub · 70 lines

Changes

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.

2mo ago First seen · 70 lines · 69 tokens per session scan A 5c07d798b2c4

Subscribe to this mod's changes

dogfood is a skill published in the GitHub repository fold-run/fold (1 stars, last pushed 12d ago), licensed Apache-2.0. It adds 69 tokens to every session and 847 once invoked, about $0.0003 per session on Opus 5.5. 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.

Debugging Go Testing ai api-gateway federation gateway go

dogfood next to 6 skills from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
dogfoodfold-run/fold	Skill	1	12d ago	Apache-2.0	A	69	Claude Code
pi-loop-forensicsdimetron/pi-go	Skill	222	3d ago	MIT	A	65	Claude Code, Codex
nightly-session-watchdimetron/pi-go	Skill	222	3d ago	MIT	A	78	Claude Code, Codex
tui-render-debugdimetron/pi-go	Skill	222	3d ago	MIT	A	33	Claude Code, Codex
kessoku-dimazrean/kessoku	Skill	143	3d ago	MIT	A	52	Claude Code, Codex
config-loader-helperpilinux/gorest	Skill	506	16d ago	MIT	A	25	Claude Code, Codex
fix-suggesterpilinux/gorest	Skill	506	16d ago	MIT	A	21	Claude Code, Codex

fold: Skill for Claude Code