ThreadShelf AGENTS.md

ThreadShelf AGENTS.md is an instructions file for Codex, OpenCode from ChrystianSchutz/ThreadShelf. It costs 4,002 tokens per session, scanned A, original, MIT.

Project instructions for ThreadShelf, an offline tool that searches and backs up exported AI chats on the user's computer.

In plain words
What is it for?
Use them when developing or reviewing ThreadShelf's chat import, local search, web interface, HTTP API, or MCP server.
Why use it?
They explain the repository layout, commands, and privacy rules for a project whose chat data is normally kept on-device.

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/chrystianschutz/threadshelf/agents-md
Clone the repo
git clone --depth 1 https://github.com/ChrystianSchutz/ThreadShelf

Made for: Codex, OpenCode.

Wrote 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.

agentmods badge for ThreadShelf AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/chrystianschutz/threadshelf/agents-md.svg)](https://agentmods.dev/instructions/chrystianschutz/threadshelf/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/chrystianschutz/threadshelf/agents-md"><img src="https://agentmods.dev/badge/instructions/chrystianschutz/threadshelf/agents-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 4,002 This file is loaded in full into every session.
When invoked 4,002 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 $0.04002 $0.04002
Opus 5 $0.02001 $0.02001
Sonnet 5 $0.00800 $0.00800
Haiku 4.5 $0.00400 $0.00400

Measured 3d ago against content hash 69c0042d37bd, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ThreadShelf 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 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.

AGENTS.md · 274 lines

How it starts

The opening of the file, as written. The whole thing — 274 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Guidance for AI coding agents (Claude Code, Codex, Cursor, etc.) working in this repository. Humans should read this too — it is the short, accurate map of the project. CLAUDE.md simply points here so Claude Code picks up the same rules.

What this project is

ThreadShelf — a local, offline semantic-search and backup tool for AI chat exports (Google AI Studio, OpenRouter, OpenAI/ChatGPT, Anthropic/Claude, LM Studio, Grok/xAI). It parses exported JSON into normalized turns, chunks and embeds them locally (Xenova/Transformers.js), stores vectors in LanceDB, and exposes search through a React web UI, an HTTP API, and an MCP stdio server.

Parsing, embeddings, storage, and search run on the user's machine. By default, no chat data leaves the device. The embedding model may be downloaded on first use. Treat all real chat exports as private.

The optional conversation-generation layer is Experimental Alpha. Its primary llama.cpp engine is local and loopback-only. OpenRouter is an explicit, opt-in external exception: picking the OpenRouter provider tab sends selected user/assistant thread content, the optional master prompt, and the new prompt off-device; archived thinking is excluded. There is no longer a per-send consent checkbox — the off-device chip on the model button and the composer hint carry that signal. New chats are saved locally by default in the protected threadshelf_conversations collection and semantic index. The explicit ghost-icon private mode is tab-scoped and never persisted.

The master prompt is a small collection of user-written system prompts. They are hand-written and expected to last, so they are stored server-side in .threadshelf/master-prompts.json (src/generation/master-prompts.ts, atomic write, mode 0600, MASTER_PROMPTS_PATH override) and served by the loopback-only /api/generation/prompts CRUD routes. The client holds no copy beyond the react-query cache (['master-prompts']). The active prompt is sent as systemPrompt on every generation request and prepended as a leading system message; it is never persisted into a stored thread, so re-reading a chat never replays a prompt the user has since changed.

Read the full file on GitHub · 274 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 · 274 lines · 4,002 tokens per session scan A 69c0042d37bd

Subscribe to this mod's changes

ThreadShelf AGENTS.md is an instructions file published in the GitHub repository ChrystianSchutz/ThreadShelf (5 stars, last pushed 12d ago), licensed MIT. It adds 4,002 tokens to every session, about $0.0200 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.

Related

Other instructions, from other repositories