copilot-memory-store copilot-instructions.md

Developer instructions for copilot-memory-store, a local JSON file that stores, searches, compresses, and removes notes for GitHub Copilot work. It also exposes these memory actions through a command line and MCP, a standard way for AI tools to call services.

In plain words
What is it for?
Changing memory storage, keyword search, compression, file locking, the interactive command line, or the MCP server in this TypeScript project.
Why use it?
They keep all storage behavior in one place and protect the command-line and MCP data formats. They also prevent concurrent writes and accidental output mixed with protocol messages.

Instructions file for GitHub Copilot

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/timothywarner-org/copilot-memory-store/copilot-instructions
Clone the repo
git clone --depth 1 https://github.com/timothywarner-org/copilot-memory-store

Made for: GitHub Copilot.

Per session 632 This file is loaded in full into every session.
When invoked 632 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 $0.00632 $0.00632
Opus 5 $0.00316 $0.00316
Sonnet 5 $0.00126 $0.00126
Haiku 4.5 $0.00063 $0.00063

Measured 3d ago against content hash 26f820435c79, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

copilot-memory-store copilot-instructions.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.

.github/copilot-instructions.md · 33 lines

How it starts

The opening of the file, as written. The whole thing — 33 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Copilot Memory Store – AI Agent Instructions

Overview

  • Local JSON-backed memory store focused on Copilot context engineering; all persistence, search, and compression flow through src/memoryStore.ts.
  • TypeScript project compiled to ESM (tsconfig targets ES2022); prefer async/await and import syntax, and avoid introducing CommonJS helpers.

Core Modules

  • src/memoryStore.ts centralizes file locking, keyword extraction, scoring, and deterministic compression—reuse these helpers instead of reimplementing I/O or search logic.
  • src/mcp-server.ts exposes memory operations as MCP tools/resources/prompts; stdout is reserved for JSON-RPC, so log via stderr with the log helper.
  • src/cli.ts provides the memory> REPL, reloading the store on each command to avoid state drift; keep CLI handlers thin wrappers over memoryStore APIs.
  • src/deepseek.ts wraps the optional DeepSeek chat-completions endpoint; it expects OpenAI-compatible responses and trims output to the requested character budget.

Memory File Behavior

  • Memories live in .copilot-memory.json by default (override with MEMORY_PATH); writes must go through addMemory/softDelete/purge so the atomic lock files remain correct.
  • Records include auto-extracted keywords (top 10, stop-word filtered) that power search scoring; preserve this contract when adding new ingestion paths.
  • Search scoring weights keyword hits, tag matches, and recency—any new ranking tweaks should adjust scoreRecord and keep CLI/MCP outputs in sync.
  • Deterministic compression builds a markdown list inside a strict character budget; respect the budget plumbing when layering new formatting.

CLI & MCP Workflows

  • CLI commands (npm run dev) rely on tokenize/parse helpers for flags like --tags or --llm; honor those conventions when extending commands.
  • MCP server runs via npm run mcp (tsx) or compiled npm run mcp:dist; any new tool must return structured text payloads compatible with MCP SDK expectations.
  • Inspector scripts (npm run inspect, npm run inspect:dev) execute the MCP server for debugging—keep transports compatible with stdio + inspector usage.

Read the full file on GitHub · 33 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. 3d ago First seen · 33 lines · 632 tokens per session scan A 26f820435c79

Subscribe to this mod's changes

copilot-memory-store copilot-instructions.md is an instructions file published in the GitHub repository timothywarner-org/copilot-memory-store (9 stars, last pushed 8mo ago), licensed MIT. It adds 632 tokens to every session, about $0.0032 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.