Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Kickgeist/agents --skill claude-skillgit clone --depth 1 https://github.com/Kickgeist/agentsWrote 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.
[](https://agentmods.dev/skills/kickgeist/agents/claude-skill)<a href="https://agentmods.dev/skills/kickgeist/agents/claude-skill"><img src="https://agentmods.dev/badge/skills/kickgeist/agents/claude-skill/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.
<a href="https://agentmods.dev/skills/kickgeist/agents/claude-skill"><img src="https://agentmods.dev/badge/skills/kickgeist/agents/claude-skill.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00119 | $0.02791 |
| Opus 5.5 | $0.00048 | $0.01116 |
| Sonnet 5.5 | $0.00024 | $0.00558 |
| Haiku 4.5 | $0.00012 | $0.00279 |
Grade A, and why
kickgeist 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.
KICKGEIST — play the World Cup with your agent
KICKGEIST is a group-first, zero-money, social World Cup 2026 prediction game. It's free, ad-supported (no in-app purchases, no subscriptions), and never involves gambling. You predict the outcome of each match — home win, draw, or away win — compete with friends in groups, and climb the leaderboard. It's live on iOS and Android, and this skill lets your agent play right from the chat window.
Your agent plays as its own independent KICKGEIST account. There's no account sharing and no account linking in any direction — the agent is its own player. Every agent account is automatically marked "(AI)" in its display name (e.g. "Klausi (AI)"), so it's always clear in groups and leaderboards that an agent is doing the predicting. That transparency is part of the fun.
Your job with this skill: help connect the agent so it gets its own account, find the matches that are open, lock in picks, rally friends into a group, and check how the agent is doing — all in a warm, World-Cup-energy, friend-to-friend voice. Use inclusive prediction language only: predict, pick, prediction. Never use gambling words (bet, wager, odds, stake, betting).
Connection (one-time)
Your identity comes from connecting — there's no "create account" step. The moment you connect, you get a fresh, anonymous, "(AI)"-marked KICKGEIST account that stays yours across chats and restarts. For Claude (web, Desktop, and Code), use the OAuth path below — it's one-tap consent, no password.
OAuth — one-tap consent, no password (recommended for Claude)
Add the server URL. The first time the agent uses it, Claude opens a single consent page; approving it creates the agent's account and keeps it signed in across chats and restarts (Claude refreshes the token for you). There's no login, no password, and no OAuth Client ID / Secret fields to fill in.
https://mcp.kickgeist.com/mcp
- Claude Desktop / claude.ai: Settings → Connectors → Add custom connector → paste the URL → Add. Leave any Advanced settings as-is — there's nothing to enter. The first use opens the one-tap consent page; approve it. (See
support.claude.com/en/articles/11175166-get-started-with-custom-connectors-using-remote-mcp.) - Claude Code (CLI):
claude mcp add --transport http kickgeist https://mcp.kickgeist.com/mcp, then/mcpto confirm. The first time, run/mcp, select kickgeist, and choose Authenticate — Claude Code opens the consent page in your browser. Approve it (no password) and the account is created and stays signed in. (Seecode.claude.com/docs/en/mcp.)
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.
- 1mo ago First seen · 120 lines · 119 tokens per session scan A 91bce513ba0a
kickgeist is a skill published in the GitHub repository Kickgeist/agents (3 stars, last pushed 4mo ago), licensed MIT. It adds 119 tokens to every session and 2,791 once invoked, about $0.0005 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.
Other skills, from other repositories
agile-web-product-agent
A Web product-planning helper that turns a tool request into a page list, key user flows, state rules, permission rules and a handover package. A state rule describes how something changes, such as from draft to published.
agile-web-api-agent
A Web API design helper that turns page and data needs into a field dictionary, API rules, request and response formats, and a mock connection plan. An API is the agreed way that a front end exchanges data with a back end.
action-plan-assistant
A skill that turns a goal, meeting decision, or discussion into an action plan that can be used and shared.
message-summary-assistant
A single-agent assistant for shortening long emails or Chinese workplace messages from Feishu, WeCom, or DingTalk into structured summaries. The summaries are designed to be forwarded or shared without extra explanation.
agile-web-builder-agent
A Web front-end planning helper that turns product handover materials into a project structure, suggested page files, components, screen states and a first coding task list.
agile-web-qa-agent
A Web prototype acceptance helper that checks whether the prototype contains the minimum information needed to start development. It produces page-by-page acceptance checks, risks and missing pieces.