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 agentmods add skills/michaeljauk/brain-starter/kickoffnpx skills add michaeljauk/brain-starter --skill kickoffgit clone --depth 1 https://github.com/michaeljauk/brain-starterWrote 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/michaeljauk/brain-starter/kickoff)<a href="https://agentmods.dev/skills/michaeljauk/brain-starter/kickoff"><img src="https://agentmods.dev/badge/skills/michaeljauk/brain-starter/kickoff.svg" alt="Measured on agentmods" 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 | $0.00080 | $0.02271 |
| Opus 5 | $0.00040 | $0.01136 |
| Sonnet 5 | $0.00016 | $0.00454 |
| Haiku 4.5 | $0.00008 | $0.00227 |
Grade A, and why
kickoff 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 5d 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 — 240 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Kickoff
Takes a freshly forked brain-starter from empty template to working vault. A preflight check, then six phases, each one independently useful. The user can stop after any phase and still have something better than they started with.
Target: 15 minutes for all six.
Trigger phrases
/kickoff/kickoff <phase>- run one phase only (identity,project,search,integrations,indexes,commit)- "I just forked brain-starter"
- "set up my brain"
- "help me personalise this vault"
Contract
These are not style preferences. Breaking one makes the skill worse than no skill.
- Every phase is gated. Ask before starting each one. The user may skip any phase, in any order, and skipping is a normal outcome - not a failure.
- Every phase verifies its own work with a tool call. An exit code of 0 is not verification. Read the file back. Run the query. Check that a known note comes out. If you did not observe the result, you may not report the phase as done.
- Never install anything without asking. Detect, report, offer the command. The user runs it, or approves you running it. This is their machine.
- Never invent facts about the user. Everything in
CLAUDE.mdandknowledge/me.mdcomes from what they told you in this conversation. An empty section beats a plausible guess. - Progress is saved after every phase to
.claude/kickoff-state.json, so an interrupted run resumes instead of restarting. - Re-running is safe. Completed phases report as already done and are skipped unless the user asks to redo one. Never overwrite the user's own edits to a file a previous phase created - read it first, and extend rather than replace.
State file
.claude/kickoff-state.json. Create it on first run. Write it after every phase.
{
"version": 1,
"started": "YYYY-MM-DD",
"phases": {
"preflight": { "status": "done", "at": "YYYY-MM-DD" },
"identity": { "status": "done", "at": "YYYY-MM-DD" },
"project": { "status": "skipped", "at": "YYYY-MM-DD" },
"search": { "status": "failed", "at": "YYYY-MM-DD", "note": "qmd not installed" },
"integrations": { "status": "pending" },
"indexes": { "status": "pending" },
"commit": { "status": "pending" }
}
}
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.
- 5d ago First seen · 240 lines · 80 tokens per session scan A 6d796e701891
kickoff is a skill published in the GitHub repository michaeljauk/brain-starter (5 stars, last pushed 9d ago), licensed MIT. It adds 80 tokens to every session and 2,271 once invoked, about $0.0004 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
log
Use this when a PersonalOS work session, chat outcome, personal reflection, or provided source should be persisted into the modular Daily context and any already-owned canonical files. Do NOT use for single-call processing, single-source knowledge ingestion, Gmail/WhatsApp propagation, or Lexware booking.
context-compression
Сжатие исторического контекста контуров и проектов: поддержка meetings/README.md как краткой памяти по прошлым встречам, решениям, разворотам, открытым вопросам, устаревшим договорённостям и якорным встречам. ОБЯЗАТЕЛЬНО используй этот скилл, когда пользователь просит: «сожми историю», «сделай скользящее окно»…
parking
Парковка текущего контекста работы для безопасного переключения на другую задачу. Агент анализирует ход беседы, определяет проект и задачу, классифицирует тип прерывания, записывает стоп-кадр в log.md проекта, обновляет plan.md (если появляется blocker или меняется вектор), а затем фиксирует состояние в правильном…
analyse-call
Use this when a local call transcript or meeting note should be analyzed into traceable PersonalOS context. Preserves the source, separates evidence from interpretation, updates only proven owner truth, routes confirmed commitments through task-manager, and never sends messages or triggers external actions without…
vault-onboarding-guide
Интерактивный проводник по первому запуску файлового хранилища с агентом. Используй, когда пользователь просит: «проведи онбординг», «помоги начать с хранилищем», «покажи методологию на практике», «хочу пощупать на реальной задаче», «научи работать с 00inbox, проектами, plan.md, tasks.md, parking/resume», «проведи по…
new-dialog-handoff
Безопасный переход в новый диалог, когда чат стал длинным или агент прогнозирует скорое исчерпание контекстного окна. Скилл переиспользует parking и resume, фиксирует durable source of truth и готовит короткий restart-packet для свежего чата. ОБЯЗАТЕЛЬНО используй, когда пользователь говорит: «новый диалог», «перенеси…