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 instructions/shofer-dev/claude-code-second-brain/agents-mdgit clone --depth 1 https://github.com/shofer-dev/claude-code-second-brainWhat 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 | $0.00513 | $0.00513 |
| Opus 5 | $0.00257 | $0.00257 |
| Sonnet 5 | $0.00103 | $0.00103 |
| Haiku 4.5 | $0.00051 | $0.00051 |
Grade A, and why
claude-code-second-brain AGENTS.md 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 yesterday.
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 — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for agents working in this repository — the Second Brain Claude
Code plugin (a background observer that drops rare, one-way advisories into a
session). The sibling CLAUDE.md is a symlink to this file.
Read first
README.md— what it is and how it works;DESIGN.md— the reasoning, including what was evaluated and rejected;PRIVACY.md— exactly what leaves the machine;TODO.md— known gaps. Keep all four in sync with the code in the same change, current-state only.
Invariants
- Silence is the success metric. Most passes must produce nothing. Any
change that makes the advisor speak more often needs an explicit,
design-level justification in
DESIGN.md— a chatty Second Brain is a broken one. - One-way and non-blocking. The plugin only reads the agent's emissions (never tool results) and only ever injects a short advisory. It must never gain the ability to block, veto, pause, edit, ask, or write to the repository — that boundary is the product.
- Privacy contract is load-bearing. Anything that changes what data is
read, sent, or stored must update
PRIVACY.mdin the same change.
Working here
- This is a public FOSS repo (github.com/shofer-dev). Never commit a credential, a private hostname, or a file referencing either by value; the plugin takes configuration at runtime.
- Run the tests with
run-tests.shand keep them green before committing. It is also the pre-push gate (git-hooks/pre-push; activate once per clone withgit-hooks/install.sh): a push with a red suite is refused. The suite needs onlypytest; the script uses.venvwhen present and bootstraps one otherwise. - Commands live in
commands/(/second-brain-*), the observer loop inworker/, hook wiring inhooks/, monitors inmonitors/— skim the directory's files before changing its behaviour.
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.
- yesterday First seen · 42 lines · 513 tokens per session scan A 53b030e7cce8
claude-code-second-brain AGENTS.md is an instructions file published in the GitHub repository shofer-dev/claude-code-second-brain (2 stars, last pushed 5d ago), licensed Apache-2.0. It adds 513 tokens to every session, about $0.0026 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 instructions, from other repositories
Cyber-AI-Forge AGENTS.md
Instructions for tanghaojie/Cyber-AI-Forge, covering ai 协作规则, 人类修改优先与冲突处理, git 安全门禁与 ai 提交标记, 开始修改前的暂存区硬门禁 and ai 自动提交的强制标记.
claudekit AGENTS.md
Instructions for duthaho/claudekit, covering agents.md, what this is, build, test, run, architecture and conventions & gotchas.
herdr-cache-alert CLAUDE.md
Instructions for AltanS/herdr-cache-alert, covering claude.md — working agreement for this repo, the claim contract — mandatory, versioning — mandatory, build / run and architecture — the four layers.
claude-bridge AGENTS.md
Instructions for kadaba/claude-bridge, covering agents.md — using claude-bridge from any ai coding cli, two independent channels, 1. coordination (the conversation), 2. file transfer (the bytes) — same for every tool and workflows.
coograph brutal-honesty.instructions.md
Global communication style — applies to all workspaces and conversations.
coograph styling.instructions.md
Conventions for CSS/styling files — loaded only when editing styles.