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/afx-team/hebb-mind/claude-mdgit clone --depth 1 https://github.com/afx-team/hebb-mindWrote 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/instructions/afx-team/hebb-mind/claude-md)<a href="https://agentmods.dev/instructions/afx-team/hebb-mind/claude-md"><img src="https://agentmods.dev/badge/instructions/afx-team/hebb-mind/claude-md.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 | $0.01046 | $0.01046 |
| Opus 5 | $0.00523 | $0.00523 |
| Sonnet 5 | $0.00209 | $0.00209 |
| Haiku 4.5 | $0.00105 | $0.00105 |
Grade A, and why
hebb-mind CLAUDE.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 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 — 93 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Hebb Mind
Neuroscience-inspired memory framework for AI agents. Open-source under github.com/afx-team, distributed on PyPI as hebb-mind.
Priority Hierarchy
When rules conflict, follow this priority order:
- MUST — Non-negotiable constraints (违反则输出无效)
- SHOULD — Strong recommendations (默认遵守，可明确说明理由后偏离)
- MAY — Optional suggestions (视具体情况采用)
Project Context
Status: Production (PyPI hebb-mind)
Domain: Long-term memory for LLM agents — write, consolidate, recall, forget
Surfaces: Python package, Click CLI (hebb), FastAPI server, MCP server (stdio), web console, Claude Code + Codex integrations
Docs site: VitePress → GitHub Pages at https://afx-team.github.io/hebb-mind/
Publication Boundary
repo_pages/is the public website (VitePress → GitHub Pages). User-facing docs go here. Never commit secrets, internal analysis, or proprietary data — anything merged is published.reports/is internal-only (papers, analysis, design, surveys). Research notes, audits, and design drafts go here, not inrepo_pages/.
Evaluation & Benchmarks
The benchmark harness lives in eval/ — datasets, metrics, CLI, and report layout are documented in eval/README.md (read it before running or updating evals). Must-know invariants:
- Each benchmark runs against an isolated server (own port
8401–8409, own workdirhebb.db) — never the dailyhebb serviceon 8321. - Match the metric to the dataset's ground truth — session
Recall@k/Hit@kwhere a clean evidence id exists; end-to-end QA only otherwise. LongMemEval QA uses the official reader +get_anscheck_promptjudge verbatim (no benchmark-tuned prompts). - A published number in
repo_pages/benchmarks/MUST cite an in-treerun-Nreport; update the EN page and itszh/mirror together.
User Path Ownership (MUST)
The user's complete path — install → first command → background operation → uninstall — is the product. Every crossing between environments (shell ↔ GUI app ↔ launchd ↔ systemd ↔ Task Scheduler ↔ a third-party CLI's subprocess) is the framework's responsibility, never the user's.
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 · 93 lines · 1,046 tokens per session scan A ffcf3be69fae
hebb-mind CLAUDE.md is an instructions file published in the GitHub repository afx-team/hebb-mind (51 stars, last pushed 1mo ago), licensed MIT. It adds 1,046 tokens to every session, about $0.0052 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.
Other instructions, from other repositories
MemOS CLAUDE.md
Claude Code instructions for MemTensor/MemOS, covering claude.md, claude code entry, sub-agents and project knowledge.
MemOS AGENTS.md
AGENTS.md instructions for MemTensor/MemOS, covering agents.md, project overview, repository layout, command cheatsheet and core api.
braindb CLAUDE.md
Claude Code instructions for dimknaf/braindb, covering braindb — claude instructions, ⚠ tool priority — read this first, it overrides habit, at the start of every session, 1. get always-on rules (behavioral guidelines) and project structure.
wayland-core copilot-instructions.md
Copilot instructions for FerroxLabs/wayland-core, covering ijfw rules, output discipline, memory routing, context discipline and cross-audit.
memory-bridge AGENTS.md
AGENTS.md instructions for jiabaobei/memory-bridge, covering agents.md — 记忆桥项目规约（ai 协作会话必读）, 版本与发布（硬性规则，所有者 2026-08-29 定）, 架构铁律 and 工程习惯.
engram CLAUDE.md
Claude Code instructions for toejough/engram, covering engram, core principles, directory structure, key files and design principles.