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 harebeats/cannonball --skill kit-promptgit clone --depth 1 https://github.com/harebeats/cannonballWrote 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/harebeats/cannonball/kit-prompt)<a href="https://agentmods.dev/skills/harebeats/cannonball/kit-prompt"><img src="https://agentmods.dev/badge/skills/harebeats/cannonball/kit-prompt.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00155 | $0.02720 |
| Opus 5 | $0.00077 | $0.01360 |
| Sonnet 5 | $0.00031 | $0.00544 |
| Haiku 4.5 | $0.00015 | $0.00272 |
Grade A, and why
kit-prompt scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 225 lines — stays where its author put it; the contents beside it link to each section on GitHub.
kit-prompt — gerar o prompt, não o site
Antes de qualquer comando: resolva o SKILL_DIR
Todo comando abaixo roda um script que viaja junto desta skill, em
SKILL_DIR/scripts/. Defina SKILL_DIR como o caminho absoluto da pasta que
contém ESTE SKILL.md que você acabou de ler — o seu harness informou esse caminho
no resultado da leitura. Funciona em qualquer hospedeiro, sem depender de variável
de ambiente de nenhum agente específico:
~/.claude/plugins/cache/cannonball/cannonball/<v>/skills/<nome>/SKILL.md
~/.codex/skills/<nome>/SKILL.md
~/.gemini/skills/<nome>/SKILL.md
~/.agents/skills/<nome>/SKILL.md
Em todos, SKILL_DIR é a pasta do SKILL.md, e SKILL_DIR/scripts/ está ao lado.
A diferença que define esta skill: kit-montar escreve código no projeto; esta
produz texto. Um prompt que você leva para outro lugar, reusa, edita, guarda.
O trabalho aqui é gerar mais um prompt no padrão dos que já funcionaram, alimentado pelas peças que já existem. Veja o que existe antes de começar:
python "${SKILL_DIR}/scripts/perfil.py"
python "${SKILL_DIR}/scripts/buscar.py" --familia prompt --qualidade favorito
Acervo vazio não impede nada aqui — o padrão abaixo é a regra, não um resumo estatístico do acervo. Só peça material ao usuário se ele quiser um prompt calcado em alguma peça específica.
O que separa um prompt bom de um mediano
Não é opinião: sai de comparar os marcados favorito contra os comuns.
| Marcador | Favorito | Comum |
|---|---|---|
| tamanho | 10.672 chars | 8.284 |
| medidas em px | 16 | 12 |
| URLs exatas | 7 | 4 |
| easings nomeados | 2 | 1 |
| proibições explícitas | 1+ | 0 |
Especificidade e restrição. Os bons dizem translateY(-40px), #0A0B11,
cubic-bezier(.16,1,.3,1) — e dizem o que não fazer. "Sem cards. Sem roxo. Não
adicione seções além das listadas." Prompt vago devolve resultado genérico.
Um prompt do acervo abaixo de ~6.000 caracteres quase sempre é rascunho. Não entregue um esqueleto e chame de pronto.
What ships with it
52 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/fundamentos-visuais.md 19 KB
- references/getlayers.md 8.5 KB
- references/higgsfield.md 4.6 KB
- references/motionsites-genjutsu-designdna.md 7.0 KB
- references/stack-animacao.md 7.1 KB
- references/stack-webgl-react.md 13 KB
- scripts/armadilhas.py 4.8 KB runs code
- scripts/assets.py 4.8 KB runs code
- scripts/buscar.py 9.4 KB runs code
- scripts/caminhos.py 6.4 KB runs code
- scripts/capturar.py 19 KB runs code
- scripts/cor.py 11 KB runs code
- scripts/curar.py 12 KB runs code
- scripts/fontes_licenca.json 19 KB
- scripts/indexar.py 10 KB runs code
- scripts/ingerir_design.py 6.7 KB runs code
- scripts/ingerir_mcp.py 7.6 KB runs code
- scripts/ingerir_projeto.py 10 KB runs code
- scripts/ingerir_registro.py 5.1 KB runs code
- scripts/ingerir.py 5.4 KB runs code
- scripts/lote.py 5.5 KB runs code
- scripts/normalizar.py 10.0 KB runs code
- scripts/painel.py 20 KB runs code
- scripts/perfil.py 10 KB runs code
- scripts/publicar.py 12 KB runs code
- scripts/receita.py 4.8 KB runs code
- scripts/relocalizar.py 6.0 KB runs code
- scripts/sinonimos.json 7.6 KB
- scripts/tags_funcao.py 5.8 KB runs code
- scripts/test_sinonimos.py 1.8 KB runs code
- scripts/tipo.py 22 KB runs code
- scripts/vincular.py 5.0 KB runs code
- seed/acervo/index.json 18 KB
- seed/acervo/ui/kit-agendamento/armadilhas.json 4.1 KB
- seed/acervo/ui/kit-agendamento/codigo/agendamento.css 8.0 KB
- seed/acervo/ui/kit-agendamento/codigo/Agendamento.tsx 13 KB
- seed/acervo/ui/kit-agendamento/codigo/regras.ts 4.8 KB runs code
- seed/acervo/ui/kit-agendamento/codigo/useAgendamento.ts 4.0 KB runs code
- seed/acervo/ui/kit-agendamento/item.json 1.7 KB
- seed/acervo/ui/kit-agendamento/README.md 6.2 KB
- seed/acervo/ui/kit-calendario/armadilhas.json 798 B
- seed/acervo/ui/kit-calendario/codigo/calendario.css 5.4 KB
- seed/acervo/ui/kit-calendario/codigo/Calendario.tsx 7.7 KB
- seed/acervo/ui/kit-calendario/codigo/useCalendario.ts 4.1 KB runs code
- seed/acervo/ui/kit-calendario/item.json 1.4 KB
- seed/acervo/ui/kit-calendario/README.md 3.9 KB
- seed/acervo/ui/kit-mapa/armadilhas.json 747 B
- seed/acervo/ui/kit-mapa/codigo/mapa.css 5.3 KB
- seed/acervo/ui/kit-mapa/codigo/Mapa.tsx 4.7 KB
- seed/acervo/ui/kit-mapa/item.json 1.6 KB
- seed/acervo/ui/kit-mapa/README.md 4.3 KB
- seed/LEIA.md 905 B
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.
- 2d ago Changed · +1 lines 214485b2df79
- 7d ago First seen · 224 lines · 155 tokens per session scan A 2995ea416955
kit-prompt is a skill published in the GitHub repository harebeats/cannonball (9 stars, last pushed 4d ago), licensed MIT. It adds 155 tokens to every session and 2,720 once invoked, about $0.0008 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 2026-08-31.
Other skills, from other repositories
stitch-ued-guide
Visual vocabulary, design terminology, and prompt engineering strategy for Stitch. Reference this when you need layout pattern names, aesthetic style terms, color structure formulas, or device guidelines.
stitch-ui-prompt-architect
Builds Stitch-ready prompts via two paths — Path A enhances vague ideas into polished prompts, Path B merges a Design Spec JSON + user request into a structured [Context] [Layout] [Components] prompt.
skill-meta-prompt
Craft better prompts using proven optimization techniques — use when your prompt needs refinement.
ai-workflow-architect
Designs AI systems, automations, and agent workflows for a business — identifying which manual work is worth automating, how to structure the system, which tools fit, and what could go wrong. Use this to automate part of an operation, design an agent or MCP workflow, reduce repetitive manual work, connect tools into a…
cloudflare-workers-ai
Cloudflare Workers AI for serverless GPU inference. Use for LLMs, text/image generation, embeddings, or encountering AIERROR, rate limits, token exceeded errors.
nano-banana-prompting
This skill should be used when crafting prompts for Nano Banana Pro (Gemini image generation). Use when users want help writing image generation prompts, need guidance on prompt structure, or want to optimize their prompts for better results.