OpenLMlib: Command for GitHub Copilot

.github/prompts/collab_sessions_plan_v2.md

collab_sessions_plan_v2 is a command for GitHub Copilot from Vedant9500/OpenLMlib. It costs 0 tokens per session (7,836 once invoked), scanned A, original, MIT.

A plan for a local system where multiple AI agents coordinate through SQLite and files. It tracks sessions, messages, and state in the database while keeping research and other work products as files.

In plain words
What is it for?
Use it to design multi-agent collaboration with session tracking, message exchange, and file-based outputs.
Why use it?
It addresses the difficulty of sharing messages, progress, and work products between several agents without relying on one central service.

Command for GitHub Copilot

Written for GitHub Copilot: a Copilot chat mode or prompt. Also seen: mentions Codex.

This is Vedant9500/OpenLMlib's own configuration. It tells GitHub Copilot how to work on OpenLMlib itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything OpenLMlib configures →

Needs its repository: it runs a file that does not travel with it, so clone the repository first. The line is openlmlib collab export <session_id> --output-dir ./research/.

Reuse

Borrowing it

Nothing to install: this file belongs to Vedant9500/OpenLMlib. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/Vedant9500/OpenLMlib/master/.github/prompts/collab_sessions_plan_v2.md
Clone the repo
git clone --depth 1 https://github.com/Vedant9500/OpenLMlib

Made for: GitHub Copilot.

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 collab_sessions_plan_v2

README.md
[![agentmods](https://agentmods.dev/badge/commands/vedant9500/openlmlib/collab_sessions_plan_v2/github.svg)](https://agentmods.dev/commands/vedant9500/openlmlib/collab_sessions_plan_v2)
Your own site
<a href="https://agentmods.dev/commands/vedant9500/openlmlib/collab_sessions_plan_v2"><img src="https://agentmods.dev/badge/commands/vedant9500/openlmlib/collab_sessions_plan_v2/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for collab_sessions_plan_v2

Your own site · 80×15
<a href="https://agentmods.dev/commands/vedant9500/openlmlib/collab_sessions_plan_v2"><img src="https://agentmods.dev/badge/commands/vedant9500/openlmlib/collab_sessions_plan_v2.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 7,836 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00000 $0.07836
Opus 5 $0.00000 $0.03918
Sonnet 5 $0.00000 $0.01567
Haiku 4.5 $0.00000 $0.00784

Measured yesterday against content hash 59d89ee3ce12, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-13, from the pricing page.

Security

Grade A, and why

collab_sessions_plan_v2 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 yesterday.

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/prompts/collab_sessions_plan_v2.md · 855 lines

How it starts

The opening of the file, as written. The whole thing — 855 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Multi-Agent Collaboration Feature: Research-Backed Implementation Plan (v2)

Executive Summary

This plan outlines CollabSessions — a local-first, hybrid SQLite+file multi-agent collaboration system for OpenLMLib. After extensive research across academic papers (ICLR 2026, arXiv), production case studies (Oracle, Google ADK, Anthropic, Zylos), and coordination pattern analysis (Tacnode, Fazm, ElectricSQL), this revised plan adopts a hybrid architecture: SQLite for the session registry, message bus, and state tracking; files for artifacts and research outputs. This converges with what the industry's best systems are doing in 2026.

Key research sources informing this plan:

  • Oracle: "Comparing File Systems and Databases for AI Agent Memory" (Feb 2026) — benchmarked FSAgent vs MemAgent
  • Zylos Research: "AI Agent Memory Architectures for Multi-Agent Systems" (Mar 2026) — framework comparison
  • Google ADK: "Architecting efficient context-aware multi-agent framework" (Dec 2025) — context engineering
  • Anthropic: "How We Built Our Multi-Agent Research System" — artifact-based memory
  • Tacnode: "8 Coordination Patterns That Actually Work" (Jan 2026) — production patterns
  • arXiv 2601.13671: "The Orchestration of Multi-Agent Systems" (Jan 2026) — academic framework

1. Core Architecture — REVISED: Hybrid Approach

1.1 Design Philosophy (Updated)

Research finding: Oracle's benchmark showed MemAgent (SQLite) beat FSAgent (files) on both latency AND quality. Zylos Research: "Shared state demands a database." But files win as an interface — LLMs already know how to use them.

Decision: Hybrid SQLite + Files

  • SQLite (collab_sessions.db): Session registry, message bus (append-only table), agent registry, state tracking. Gives us ACID transactions, FTS5 search, row-level locking, zero platform-specific code.
  • Files (sessions/{id}/artifacts/): Research outputs, agent working notes, large artifacts. LLM-native interface, human-readable, versionable.
  • JSONL shadow log (sessions/{id}/messages.jsonl): Human-readable copy of the messages table for debugging. Written alongside SQLite inserts.

Read the full file on GitHub · 855 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. yesterday First seen · 855 lines · 0 tokens per session scan A 59d89ee3ce12

Subscribe to this mod's changes

collab_sessions_plan_v2 is a command published in the GitHub repository Vedant9500/OpenLMlib (2 stars, last pushed 2d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 7,836 tokens. 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 2026-09-11.