agentmods.dev

memXT AGENTS.md

Instructions for memXT, a local persistent memory system shared by several coding-agent tools. It stores project facts, decisions, documents, and past work on the local machine.

In plain words
What is it for?
Use them at the start of sessions, before answering questions about earlier work, and after making important project decisions.
Why use it?
They explain when to load prior context, search memories, expand relevant results, and save decisions so agents do not repeatedly lose project knowledge.

Instructions file for CodexOpenCode

23 1mo ago A scan 379 tokens original MIT
View the source file Yupcha/memXT
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/yupcha/memxt/agents-md
Clone the repo
git clone --depth 1 https://github.com/Yupcha/memXT

Made for: Codex, OpenCode.

Per session 379 This file is loaded in full into every session.
When invoked 379 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 $0.00379 $0.00379
Opus 5 $0.00189 $0.00189
Sonnet 5 $0.00076 $0.00076
Haiku 4.5 $0.00038 $0.00038

Measured 2d ago against content hash 07a7bfbf73c4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

memXT 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 2d 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.

AGENTS.md · 31 lines

What it actually says

Agent instructions (memxt)

Memory (memxt)

You have a persistent, local memory palace via the memory MCP server. Nothing leaves this machine. Shared across Claude Code / Codex / Cursor / Grok CLI. No cloud LLM is used to compress memory.

Progressive disclosure (default)

  1. memory_search — index by default (compact #id + snippet)
  2. memory_get — full verbatim body for promising id / ids only
  3. Prefer detail: "full" only when you need every body immediately

Tools

  • memory_wake_up — call at the start of every session (Grok does not auto-inject)
  • memory_profile — stable project facts (fast, no embedding model)
  • memory_search — before answering about prior work; modes: hybrid | memories | documents | facts | episodes
  • memory_get — expand index hits by drawer id
  • memory_store — after a real decision; prefer room: "decisions"
  • memory_forget — remove wrong/stale memories by id
  • memory_dream — optional consolidation (hot budget + 4-bit quant)
  • memory_stats — health

Rules

  1. At session start: memory_wake_up then continue.
  2. Before "what did we decide / how does X work here": memory_search → memory_get.
  3. After a real decision or user correction: memory_store one crisp fact.
  4. Default wing for this repo is memxt (or git-derived).
  5. Claude Code hooks also autosave session tails (Stop + PreCompact) locally — still store decisions explicitly.
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. 2d ago First seen · 31 lines · 379 tokens per session scan A 07a7bfbf73c4

Subscribe to this mod's changes

memXT AGENTS.md is an instructions file published in the GitHub repository Yupcha/memXT (23 stars, last pushed 1mo ago), licensed MIT. It adds 379 tokens to every session, about $0.0019 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 2026-08-30.

Memory and contextagent-memoryai-memoryclaude-codemcpzig
Related

Other instructions, from other repositories

neo AGENTS.md

AGENTS.md instructions for neomjs/neo, covering ai agent per-turn operational mandates, §corevalues, §identitypromptfirewall, §criticalgates and §precommitgates.

neomjs/neo · 5,647 tokens

plur CLAUDE.md

Instructions for plur-ai/plur, covering claude.md, what is plur, development, package dependency and version bumps.

plur-ai/plur · 4,509 tokens

dense-mem AGENTS.md

Instructions for markhuangai/dense-mem, covering dense-mem repository guidance, project context, architecture decision records, current stack and target architecture.

markhuangai/dense-mem · 2,447 tokens

wife AGENTS.md

Instructions for ma-nucho-pro/wife, covering if the user wants to install it, if the user wants to change something and rules that are not negotiable.

ma-nucho-pro/wife · 676 tokens

dakera-cli CLAUDE.md

Instructions for Dakera-AI/dakera-cli, covering dakera-cli, key commands, architecture and conventions.

Dakera-AI/dakera-cli · 282 tokens

sigil CLAUDE.md

Instructions for Anmol-Srv/sigil, covering sigil and design system.

Anmol-Srv/sigil · 128 tokens