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/notsointresting/mnemex/claude-mdgit clone --depth 1 https://github.com/notsointresting/mnemexWrote 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/notsointresting/mnemex/claude-md)<a href="https://agentmods.dev/instructions/notsointresting/mnemex/claude-md"><img src="https://agentmods.dev/badge/instructions/notsointresting/mnemex/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 | $0.00441 | $0.00441 |
| Opus 5 | $0.00220 | $0.00220 |
| Sonnet 5 | $0.00088 | $0.00088 |
| Haiku 4.5 | $0.00044 | $0.00044 |
Grade A, and why
mnemex 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 3d 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.
What it actually says
mnemex — Build Context
Goal
MCP server giving coding agents ANCHORED memory: every decision is bound to a file+symbol and hash-stamped, delivered JUST-IN-TIME via hooks. One SQLite file (sqlite-vec + FTS5). Distributed via MCP + skills + AGENTS.md across all agents.
Orchestration rules
- Opus 4.8 orchestrates; delegate bounded deliverables to subagents with written contracts.
- Subagents return diff + test results + self-report. Never blind-merge.
- A phase is done ONLY when its verification gate passes (tests green + verifier's adversarial cases pass + objective metric met).
- verifier-agent must be a SEPARATE spawn from the implementer. Authors don't grade themselves.
- Build the ANCHOR core (Phase 1) before any indexer, using synthetic nodes.
Hard constraints
- No cloud/external API for core features. Local-first.
- Context injections respect hard token caps (800 session / 400 JIT). Never exceed.
- pip install works on macOS/Linux/Windows, Py 3.10–3.13, no ML model required (no-ML mode).
- Secrets/PII stripped at write time; Phase 6 security gate is blocking.
- Prefer wrapping an existing structural backend over reinventing a parser (Path A).
Layout
mnemex/storage.py → sqlite-vec + FTS5, schema, migrations mnemex/anchors.py → anchor resolve + hash stamp + staleness (THE MOAT) mnemex/retrieval.py → BM25+vector RRF + token governor mnemex/indexer.py → structural backend adapter mnemex/server.py → FastMCP tools mnemex/hooks.py → SessionStart/PreToolUse/Stop, JIT injection mnemex/agents_md.py → self-updating AGENTS.md mnemex/security.py → secret stripping + audit log tests/fixtures/ → golden repo, Q→memory set, secret corpus, git history
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.
- 3d ago First seen · 33 lines · 441 tokens per session scan A 83454557e3d7
mnemex CLAUDE.md is an instructions file published in the GitHub repository notsointresting/mnemex (1 stars, last pushed 1mo ago), licensed MIT. It adds 441 tokens to every session, about $0.0022 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
open-skill-sunset AGENTS.md
AGENTS.md instructions for ooocooc/open-skill-sunset, a project described as: Local, read-only audit for stale AGENTS.md, CLAUDE.md, and generic SKILL.md instructions.
anchor copilot-instructions.md
Instructions for peva3/anchor, covering github copilot instructions, required reading, critical rules, prohibited actions and project structure.
anchor CLAUDE.md
Instructions for peva3/anchor: All project rules, patterns, and requirements are defined in AGENTS.md.
anchor AGENTS.md
Instructions for peva3/anchor, covering agents.md — anchor, how to use this file, section index (glossary), quick-navigation cheatsheet and core rules (kept in full — always apply).
dev-agent-skills CLAUDE.md
Instructions for fvadicamo/dev-agent-skills, covering claude.md, project overview, architecture, key file: marketplace.json and skill anatomy.
scopeglass AGENTS.md
AGENTS.md instructions for zackabrah/scopeglass, covering scopeglass contributor instructions, project, commands, conventions and boundaries.