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/feng111222-113/mcp-memory/agents-mdgit clone --depth 1 https://github.com/feng111222-113/mcp-memoryWrote 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/feng111222-113/mcp-memory/agents-md)<a href="https://agentmods.dev/instructions/feng111222-113/mcp-memory/agents-md"><img src="https://agentmods.dev/badge/instructions/feng111222-113/mcp-memory/agents-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 | $0.01645 | $0.01645 |
| Opus 5 | $0.00822 | $0.00822 |
| Sonnet 5 | $0.00329 | $0.00329 |
| Haiku 4.5 | $0.00164 | $0.00164 |
Grade A, and why
mcp-memory AGENTS.md scanned grade A 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.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
- **Windows test quirk**: if `bun` is installed via npm (`node_modules/bun/bin/bun.exe`), `child_process.spawn("bun")` fails because it's a `.cmd` wrapper. Tests set `shell: true` on Windows to work around this. How it starts
The opening of the file, as written. The whole thing — 128 lines — stays where its author put it; the contents beside it link to each section on GitHub.
mcp-memory — Agent Guide
Bun-only MCP server for persistent memory across OpenCode sessions. Markdown files on disk + SQLite FTS5 search.
Runtime & Build
- Bun only — uses
bun:sqlite, runs.tsdirectly. NOT Node.js-compatible. tsconfig.jsonhasnoEmit: true; no compile step needed.- Only dependency:
@modelcontextprotocol/sdk.
Key Commands
bun install # install
bun run src/index.ts # dev / run directly
bunx @mimochamber/memory-server # run published version
Tests (all are child-process integration tests over JSON-RPC stdio)
bun run test.mjs # smoke test
bun run test-dedup.mjs # dedup test
bun run test-compaction.mjs # compaction append test
- Tests spawn the server as a child process via
bun run src/index.ts. - All set
MCP_MEMORY_POLL_INTERVAL=0to disable compaction polling. - Use
.test-memory/temp dir (in.gitignore), cleaned up after run. - No test framework — plain
.mjsscripts, JSON-RPC over stdin/stdout. - There is no CI pipeline (no
.github/workflows/).
Publishing
.\scripts\publish.ps1 # dry-run (npm pack --dry-run)
.\scripts\publish.ps1 -Execute # npm publish --access public
Architecture
Single module, 5 source files:
src/
index.ts — MCP server setup, tool handlers, compaction poller
store.ts — MemoryStore: SQLite DB + file ops (core logic)
fts-query.ts — Builds FTS5 MATCH expression from user query (OR-join tokens)
paths.ts — Memory directory layout, path parsing/building, safety checks
types.ts — TypeScript types
bin/
mcp-memory.js — npm binary entry (chdirs to pkg dir, imports src/index.ts)
Entry point
- Dev:
src/index.ts— imports directly, Bun runs it. - Production:
bin/mcp-memory.js— changes CWD to package root, thenimport("../src/index.ts").
MemoryStore (src/store.ts)
- SQLite DB (
memory.db) with FTS5 virtual table in external-content mode. - FTS5 tokenizer:
porter unicode61(does NOT split CJK individually by default). - CJK preprocessing:
addCjkSpacing()inserts spaces after every CJK character before FTS5 indexing, so each CJK character becomes a separate FTS5 token.removeCjkSpacing()reverses this on search results. Both query and content go through this pretreatment. - Triggers auto-sync
memory_ftstable ↔ FTS5 index on INSERT/UPDATE/DELETE. - Fingerprint =
${size}-${mtimeMs}for incremental reconcile. - Path safety:
assertSafeComponentrejects.., null bytes, Windows absolute paths. - Write safety: per-file
writeQueueserializes concurrent compaction appends.
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 · 128 lines · 1,645 tokens per session scan A 8626b2b2a7b3
mcp-memory AGENTS.md is an instructions file published in the GitHub repository feng111222-113/mcp-memory (0 stars, last pushed 2mo ago), licensed MIT. It adds 1,645 tokens to every session, about $0.0082 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
aeon CLAUDE.md
Instructions for aeonfun/aeon, covering aeon, how aeon works, strategy, voice and soul file hierarchy (read in this order).
cognirepo CLAUDE.md
Claude Code instructions for ashlesh-t/cognirepo, covering claude.md, key rules, session start sequence (run in this order), behavioral confirmation rule and personas (cognirepo-402, cognirepo-403).
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.