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/evan-moon/memex/claude-mdgit clone --depth 1 https://github.com/evan-moon/memexWrote 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/evan-moon/memex/claude-md)<a href="https://agentmods.dev/instructions/evan-moon/memex/claude-md"><img src="https://agentmods.dev/badge/instructions/evan-moon/memex/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 | $0.02539 | $0.02539 |
| Opus 5 | $0.01269 | $0.01269 |
| Sonnet 5 | $0.00508 | $0.00508 |
| Haiku 4.5 | $0.00254 | $0.00254 |
Grade A, and why
memex 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 today.
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 — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Memex
Memex
Evan's second brain — semantic search over personal notes, powered by SQLite + vector embeddings.
Architecture
apps/
├── cli/ # memex CLI — the safety net (see surface policy below)
├── desktop/ # The Electron app — the only way a person reaches the screen
├── mcp/ # MCP server — exposes memex tools to Claude
└── ui/ # The React screens, served by the app's own `memex://` handler
packages/
├── core/ # Note service shared by cli/mcp (save/edit/search/delete)
├── db/ # SQLite client, schema, repository (drizzle + sqlite-vec)
├── embed/ # Embedding model wrapper (@huggingface/transformers)
├── rerank/ # Cross-encoder reranker, opt-in via MEMEX_RERANK=1
└── utils/ # Config loader, path helpers, formatters, note chunker
DB lives at ~/.memex/memex.db. Model cache at ~/.memex/models/.
openDb snapshots the database to memex.db.bak-<YYYYMMDD-HHMMSS> whenever a schema migration
is pending and the vault is not empty, then keeps the newest three. It uses VACUUM INTO rather
than copying the file, because the app, the MCP server and the CLI are all attached in WAL mode
and a filesystem copy can be torn mid-transaction — if you ever copy the DB by hand, take
-wal with it. A snapshot that fails does not stop the database from opening; memex stats
prints how many backups are kept and when the newest was taken.
Working on the UI
yarn dev:app runs the app against your real vault with the screen hot-reloading: Vite owns the
page and swaps a component without losing where you were, Electron owns everything else. The window
still loads memex://app/ — the handler asks Vite for the page instead of reading it off disk, so
/api goes the same way it does in a packaged build. yarn desktop is the same thing without Vite,
which is what to reach for when the question is about packaging rather than the screen.
better-sqlite3 is built against V8's ABI, so one build cannot serve both runtimes. yarn native
keeps two — better_sqlite3-node.node and better_sqlite3-electron.node, side by side in
node_modules/better-sqlite3/build/Release — and openDb names the one it can load. The app, the
MCP server and the tests all run at the same time; nothing is swapped, and nothing has to be put
back. dev:app, desktop and package:mac run it for you, and it does nothing when both builds
are already there for the current Node and Electron.
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.
- today Changed · +32 lines · +605 tokens per session dc90ad4774dd
- 4d ago First seen · 121 lines · 1,934 tokens per session scan A 6b68d1636240
memex CLAUDE.md is an instructions file published in the GitHub repository evan-moon/memex (3 stars, last pushed today), licensed MIT. It adds 2,539 tokens to every session, about $0.0127 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
siyuan AGENTS.md
AGENTS.md instructions for siyuan-note/siyuan, covering agents.md, 1. non-negotiable constraints, do not hand-edit, verification and prohibited operations and 2. project-specific rules.
MCP-memento copilot-instructions.md
Instructions for caiowilson/MCP-memento, covering mcp tooling policy (vs code), project architecture (big picture), critical workflows, project-specific conventions and tool notes.
vecgrep CLAUDE.md
Instructions for jeffbai996/vecgrep, covering vecgrep — claude code working notes, what this is, architecture cheat sheet, design rules and don'ts.
cerebellum CLAUDE.md
Instructions for jj-valentine/cerebellum, covering cerebellum, what this is, stack, commands and key files.
commonplace CLAUDE.md
Instructions for briannaworkman/commonplace: This repository holds two vaults. Nothing should be ingested or edited here at the root.
lien review.instructions.md
Instructions for getlien/lien, covering lien code review instructions, ⚠️ critical: lien-specific issues, no-data honesty (packages/cli/src/utils/scan-failure.ts), deterministic over inferred (packages/parser/src/signals/) and published-package surface (packages/parser/src/index.ts).