memory-star AGENTS.md

An AGENTS.md instruction file that tells an AI coding assistant how to use MemoryStar, a local database for storing and searching project knowledge.

In plain words
What is it for?
Use it to guide automatic saving and retrieval of project knowledge, including design decisions, workarounds, and code patterns.
Why use it?
It helps preserve architecture notes, conventions, decisions, dependencies, bugs, and other project information across sessions.

Instructions file for CodexOpenCode

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/ussdeveloper/memory-star/agents-md
Clone the repo
git clone --depth 1 https://github.com/ussdeveloper/memory-star

Made for: Codex, OpenCode.

Per session 1,623 This file is loaded in full into every session.
When invoked 1,623 The same file — it is already loaded in full.
Security scan A 0 findings. 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 $0.01623 $0.01623
Opus 5 $0.00812 $0.00812
Sonnet 5 $0.00325 $0.00325
Haiku 4.5 $0.00162 $0.00162

Measured today against content hash 3cf009426a7a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

memory-star AGENTS.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.

AGENTS.md · 178 lines

How it starts

The opening of the file, as written. The whole thing — 178 lines — stays where its author put it; the contents beside it link to each section on GitHub.

MemoryStar Agent Instructions

Place this file as .instructions.md or AGENTS.md in your project,

or add it to your agent's system prompt.

The agent will automatically use MemoryStar tools to manage

project knowledge.

MCP Tools - MemoryStar

You are connected to the MemoryStar MCP server - a local semantic memory database for this project. Use the tools below to save and search project information.

When to save (memory_save)

  • When you discover a new pattern, convention, or structure in the code
  • After analyzing architecture - save its description
  • When you encounter non-standard dependencies or configurations
  • After making a design decision - save as ADR (scope: "decisions")
  • When you find a bug or workaround - scope: "bugs"
  • CRITICAL: When your context usage approaches 85%, call memory_context_check
    • the server will auto-save a summary

How to save

Use descriptive key and appropriate scope:

scope = "architecture"   -> directory structure, layers, components
scope = "api"            -> endpoints, controllers, GraphQL, REST
scope = "dependencies"   -> libraries, versions, compatibility
scope = "patterns"       -> code patterns, naming conventions
scope = "decisions"      -> architectural decisions (ADR)
scope = "bugs"           -> known bugs, workarounds
scope = "notes"          -> general notes

Searching (memory_search)

Always search MemoryStar before starting work on a new task to check if notes already exist on the topic.

Use mode: "semantic" for conceptual queries (e.g. "how does auth work"), and mode: "text" for specific keywords (e.g. "JWT token middleware").

Linking (memory_link)

Create relationships between notes to build a knowledge graph:

  • depends_on - component A depends on B
  • parent_of - parent module
  • references - loose association
  • implements - interface/pattern implementation

Context Monitoring

Periodically check memory_context_check(context_usage_pct=..., project="your-project-name"). The server will assess whether a summary needs to be saved.

Read the full file on GitHub · 178 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. today First seen · 178 lines · 1,623 tokens per session scan A 3cf009426a7a

Subscribe to this mod's changes

memory-star AGENTS.md is an instructions file published in the GitHub repository ussdeveloper/memory-star (0 stars, last pushed 2mo ago), licensed MIT. It adds 1,623 tokens to every session, about $0.0081 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.