Lorekeeper CLAUDE.md

Lorekeeper CLAUDE.md is an instructions file for coding agents from Jessinra/Lorekeeper. It costs 5,195 tokens per session, scanned B, original, Apache-2.0.

Repository instructions for Lorekeeper, a personal AI memory server that stores and searches information and updates its own agent guidance over time. It uses a six-layer design with one-way dependencies between layers.

In plain words
What is it for?
Use them when changing Lorekeeper’s memory tools, data storage, server interfaces, agent configuration, or code that must follow the six-layer import rules.
Why use it?
They preserve the project’s architecture and explain how memory features and self-updating development processes are expected to work.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/jessinra/lorekeeper/claude-md
Clone the repo
git clone --depth 1 https://github.com/Jessinra/Lorekeeper

Wrote 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.

agentmods badge for Lorekeeper CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/jessinra/lorekeeper/claude-md.svg)](https://agentmods.dev/instructions/jessinra/lorekeeper/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/jessinra/lorekeeper/claude-md"><img src="https://agentmods.dev/badge/instructions/jessinra/lorekeeper/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 5,195 This file is loaded in full into every session.
When invoked 5,195 The same file — it is already loaded in full.
Security scan B 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.05195 $0.05195
Opus 5 $0.02597 $0.02597
Sonnet 5 $0.01039 $0.01039
Haiku 4.5 $0.00519 $0.00519

Measured 4d ago against content hash 7b079fa7da1b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

Lorekeeper CLAUDE.md scanned grade B with 1 finding 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 4d 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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

17. Update `~/.claude/settings.json` → restart Claude Code
CLAUDE.md · 394 lines

How it starts

The opening of the file, as written. The whole thing — 394 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Lorekeeper v2 — Living Agentic Loop

What this is: A personal AI memory MCP server, and simultaneously the first working sample of the Living Agentic Loop — a self-improving agent system that autonomously updates its own CLAUDE.md, skills, and code over time by capturing session learnings and feeding them back in.

The repo serves two purposes:

  1. The product: MCP server providing lore_search, lore_remember, lore_insert, lore_update, lore_forget, lore_recommend_links. Replaces the Node.js v1 with Python + LanceDB.
  2. The demonstration: The development process itself is looped. Session learnings are captured → consolidated → applied back to agent config. This repo is the proof of concept.

Data dir: ~/.lorekeeper (LanceDB + SQLite; controlled by LORE_DATA_DIR)


Architecture — Six-Layer (LKPR-105)

Lorekeeper follows a strict 6-layer architecture with unidirectional imports — a layer may import from layers below it, never above.

┌──────────────────────────────────────────────────────────────┐
│  api/mcp/, dashboard/, server.py, cli/                       │ Layer 6 — Interface adapters
├──────────────────────────────────────────────────────────────┤
│  shared/ (serializers, encouragement)                        │ Layer 5 — Presentation helpers
├──────────────────────────────────────────────────────────────┤
│  processors/ (memory, link, reflection, suggestion, admin)   │ Layer 4 — Orchestration
├──────────────────────────────────────────────────────────────┤
│  domains/{memory,link,reflection,suggestion}/                │ Layer 3 — Business logic
├──────────────────────────────────────────────────────────────┤
│  platform/{config,metrics}/                                  │ Layer 2 — Supporting repos
├──────────────────────────────────────────────────────────────┤
│  infra/ (database, search_engine, keyword_index,             │ Layer 1 — Zero business
│         scheduler, logging_setup, settings)                  │           vocabulary
└──────────────────────────────────────────────────────────────┘
```text

Read the full file on GitHub · 394 lines

Changes

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.

  1. 4d ago First seen · 394 lines · 5,195 tokens per session scan B 7b079fa7da1b

Subscribe to this mod's changes

Lorekeeper CLAUDE.md is an instructions file published in the GitHub repository Jessinra/Lorekeeper (4 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 5,195 tokens to every session, about $0.0260 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.