gini-agent: Skill for Claude Code

.claude/skills/dogfood-as-user/SKILL.md

dogfood-as-user is a skill for Claude Code from Open-Curiosity/gini-agent. It costs 96 tokens per session (968 once invoked), scanned A, original, MIT.

A testing method that checks an agent by sending it a normal user request in a real chat, without telling it which behavior to use. Dogfooding means using a product as an ordinary user would.

In plain words
What is it for?
Use it to verify changes to instructions, tools, dispatch, providers, memory, or skills by observing the agent handle a realistic chat turn.
Why use it?
It shows whether an agent change works by default, rather than only succeeding because the test prompt gives it coaching.

Skill for Claude Code

Written for Claude Code: installed under .claude/.

This is Open-Curiosity/gini-agent's own configuration. It tells Claude Code how to work on gini-agent itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything gini-agent configures →

About the project

Gini Agent is a personal AI-agent runtime that manages conversations, runs, tasks, approvals, memory, skills, jobs, tools, traces, audit events, and runtime health across its clients. It is for an agent that can remember, learn from task outcomes, and operate through desktop or mobile interfaces with user controls. Catalogue add-ons extend the runtime with skills and instructions.

Open-Curiosity/gini-agent · 2,142 stars · on GitHub

Reuse

Borrowing it

Nothing to install: this file belongs to Open-Curiosity/gini-agent. 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/Open-Curiosity/gini-agent/main/.claude/skills/dogfood-as-user/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/Open-Curiosity/gini-agent

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.

agentmods badge for dogfood-as-user

README.md
[![agentmods](https://agentmods.dev/badge/skills/open-curiosity/gini-agent/dogfood-as-user/github.svg)](https://agentmods.dev/skills/open-curiosity/gini-agent/dogfood-as-user)
Your own site
<a href="https://agentmods.dev/skills/open-curiosity/gini-agent/dogfood-as-user"><img src="https://agentmods.dev/badge/skills/open-curiosity/gini-agent/dogfood-as-user/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.

agentmods 80×15 button for dogfood-as-user

Your own site · 80×15
<a href="https://agentmods.dev/skills/open-curiosity/gini-agent/dogfood-as-user"><img src="https://agentmods.dev/badge/skills/open-curiosity/gini-agent/dogfood-as-user.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 96 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 968 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. Third-party audits
  • NVIDIA SkillSpector warn 7 Sept 2026
SkillSpector: 1 finding, up to medium

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

  • medium Rogue Agent · line 26
    Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
    Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
How audits are shown
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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00096 $0.00968
Opus 5 $0.00048 $0.00484
Sonnet 5 $0.00019 $0.00194
Haiku 4.5 $0.00010 $0.00097

Measured 12d ago against content hash fa1f1c118c4f, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

dogfood-as-user 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 12d 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-as-user/SKILL.md · 42 lines

How it starts

The opening of the file, as written. The whole thing — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Dogfood Gini as a user

When you change agent behavior — a steer in INSTRUCTIONS.md, a tool, dispatch, providers, memory, or skill wiring — the only real test is a real chat turn driven as a real user. Unit tests verify the mechanism; the chat turn verifies the model actually reaches for it.

The one rule: bare, uncoached prompts

Send exactly what a real user would type — and nothing more. Never narrate the intended behavior into the message.

  • Buy me a one-day fishing license day pass for California.
  • Buy me a fishing license. Drive the purchase as far as you can in the browser before involving me.
  • ... use your handoff flow / ... ask me with a choice card / ... do as much as possible without me

A coached prompt tests instruction-following, not the default the change is meant to install — and it routinely makes a behavior look more robust than it is, even producing a structured affordance (e.g. an ask_user choice card) that the bare prompt never triggers. The behavior belongs in INSTRUCTIONS.md, never in the user's mouth.

Proven here: the same task, coached ("drive as far as you can before involving me"), produced an ask_user card and a browser handoff; the bare prompt only described the options in prose and ended the turn. The coaching masked a real gap. Always send the bare request, then judge whether the agent gets there on its own.

Procedure

  1. Instance — use the worktree's own instance (the basename of the workspace dir), never default.
  2. Gateway uptmux new-session -d -A -s gini-<instance> "bun run gini run --instance <instance>"; confirm with gini status --instance <instance> (look for "ok": true).
  3. Fresh session — create a new chat/agent so no earlier coaching is sitting in context.
  4. Send the bare request the way the user does — type it into the web chat UI in a browser (agent-browser, so clientSurface is web), or mobile for a mobile change. One message, no scaffolding. The gini chat send <session> "<prompt>" CLI is a fallback for headless contexts only: it posts to the same /api/chat/<id>/messages endpoint but skips the surface the user actually touches, so it can't catch a client-side rendering regression.
  5. Observe what the user sees first — watch the response render in the web UI (screenshot to eyeball it). The task's recentToolCalls and ~/.gini/instances/<instance>/logs/runtime.jsonl are a supplement for confirming which tool fired, not the verdict — client-side rendering (chips, narration folding, cards) never shows up there. Judge whether the agent reaches the intended behavior / selects the right tool / emits the right structured affordance unprompted.
  6. Judge honestly — success is getting there on its own. If it only gets there when coached, that's a FAIL of the change, not a pass — say so plainly and quote what it actually did.

Read the full file on GitHub · 42 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.

  1. 12d ago First seen · 42 lines · 96 tokens per session scan A fa1f1c118c4f

Subscribe to this mod's changes

dogfood-as-user is a skill published in the GitHub repository Open-Curiosity/gini-agent (2,142 stars, last pushed 1mo ago), licensed MIT. It adds 96 tokens to every session and 968 once invoked, about $0.0005 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 2026-08-30.

Related

Other skills, from other repositories

reproduce-bug

Reproduce a reported bug in googleapis/mcp-toolbox and decide whether it is real, delivering an evidence-backed verdict: confirmed, already fixed, misconfiguration, client-side, works as intended, not reproducible, or blocked. Use whenever a maintainer asks you to reproduce, verify, confirm, or investigate a bug…

googleapis/mcp-toolbox · 130 tokens

langbot-testing

Test LangBot WebUI and core product flows with an automated browser and backend logs. Use when validating the configured LangBot frontend, pipeline Debug Chat, model provider setup and test buttons, bot and knowledge-base UI flows, or troubleshooting failed LangBot end-to-end tests.

langbot-app/LangBot · 58 tokens

app_verification

Verify a newly created or newly installed Rome app after coding:appcreation or coding:workflowcreation finishes installing it. Use only as a second-pass verifier: receive the app id, source root, expected behavior, and local dashboard/API base URL from the creator, then visit the installed app, exercise its safe…

rome-os/rome · 0 tokens

local-test

Build, run, and test IronClaw locally using Docker containers and Chrome MCP browser automation.

suyoumo/ClawProBench · 22 tokens

button-check

Traverse all interactive elements on the page — click clickables.

MigoXLab/webqa-agent · 15 tokens

ui-data-testid

Add stable data-testid attributes by default for new or refactored UI components. Use when implementing React/TSX views, shadcn/antd-style components, dropdown/menu configs, or interactive UI flows that need reliable selectors for unit/E2E tests.

dtyq/magic · 58 tokens