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/raskolnikovdd/adaptive-memory-graph/claude-mdgit clone --depth 1 https://github.com/raskolnikovdd/adaptive-memory-graphWrote 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/raskolnikovdd/adaptive-memory-graph/claude-md)<a href="https://agentmods.dev/instructions/raskolnikovdd/adaptive-memory-graph/claude-md"><img src="https://agentmods.dev/badge/instructions/raskolnikovdd/adaptive-memory-graph/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 | $0.00484 | $0.00484 |
| Opus 5 | $0.00242 | $0.00242 |
| Sonnet 5 | $0.00097 | $0.00097 |
| Haiku 4.5 | $0.00048 | $0.00048 |
Grade A, and why
adaptive-memory-graph 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 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.
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 — 38 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Adaptive Memory Graph — Runtime Instructions
This project is an MCP server plugin that provides Claude with persistent, intelligent memory across sessions via a weighted graph of interconnected knowledge nodes.
Runtime Behaviour
At the start of each conversation, call amg_load_index() to receive a lightweight map of the user's memory graph. Use this map as background awareness only — do not reference it unless contextually relevant.
If the conversation touches a domain or topic that clearly warrants deeper context, call amg_expand_branch() for the relevant node. Be conservative: only expand when it would genuinely improve your response.
Do not surface personal nodes during professional work sessions, and vice versa, unless explicitly relevant.
At the end of each conversation, call amg_log_session() with a summary of which branches were accessed, which were engaged with, and any suggested new nodes. Do this silently — do not narrate this process to the user.
If the user explicitly corrects a memory ("that's not relevant" / "stop bringing that up"), record this as an explicit correction in the session log.
Available Tools
amg_load_index— Load lightweight index at session startamg_expand_branch— Fetch full node content when contextually relevantamg_get_connected_nodes— Find related nodes across domainsamg_log_session— Log session summary at conversation endamg_update_graph— Process pending logs and apply decayamg_export_report— Generate human-readable graph summaryamg_manual_adjust— Boost, decay, archive, or delete nodesamg_add_node— Manually add new nodes to the graphamg_search_nodes— Search nodes by title, summary, tags, or content
Project Structure
src/graph.py— Core graph data model (Node, Graph, SessionLog)src/crypto.py— AES-256-GCM encryption with macOS Keychain key storagesrc/storage.py— Storage abstraction (LocalStorageBackend)src/update.py— Weight update, decay, and pruning logicsrc/server.py— MCP server entry point with all tool definitionsmigrations/v1_0.py— Schema migration scriptstests/test_graph.py— Test suite
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.
- 4d ago First seen · 38 lines · 484 tokens per session scan A 967f5d0797d4
adaptive-memory-graph CLAUDE.md is an instructions file published in the GitHub repository raskolnikovdd/adaptive-memory-graph (0 stars, last pushed 5mo ago), licensed MIT. It adds 484 tokens to every session, about $0.0024 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
vespertide AGENTS.md
AGENTS.md instructions for dev-five-git/vespertide, covering vespertide knowledge base, structure, where to look, data flow and conventions.
aeon CLAUDE.md
Instructions for aeonfun/aeon, covering aeon, how aeon works, strategy, voice and soul file hierarchy (read in this order).
wayland-core copilot-instructions.md
Copilot instructions for FerroxLabs/wayland-core, covering ijfw rules, output discipline, memory routing, context discipline and cross-audit.
mcp-structured-memory CLAUDE.md
Claude Code instructions for nmeierpolys/mcp-structured-memory, a project described as: Structured Memory MCP Server.
inkwell-memory CLAUDE.md
Instructions for veronchenko/inkwell-memory, covering claude.md — inkwellmemory, layout, multi-tenant mode (inkwellmultitenant=1), conventions and testing.
RNR-Enhanced-Cognee AGENTS.md
AGENTS.md instructions for vincentspereira/RNR-Enhanced-Cognee, covering rnr enhanced cognee implementation for codex, critical requirements, 1. ascii-only output (no unicode encoding), 2. dynamic categories (no hardcoded categories) and 3. standard memory mcp interface.