This repository was archived by the owner on Jun 19, 2026. It is now read-only.
Releases: sanic732/P2P
Releases · sanic732/P2P
Release list
P2P v1.1 — Claude Edition
P2P v1.1 — Claude Edition
Meta-prompt system for generating, adapting and cross-model porting of prompts across Claude, Gemini, GPT, DeepSeek, Qwen, Kimi.
Archive contents (P2P-v1.1.zip)
22 files · 20 system modules + full docs:
README.md— getting started, architecture overviewCHANGELOG.md— version history (v0.9 → v1.0 → v1.1)!!core_claude.md,!!db_claude.md— critical core (always loaded)!agents_claude.md— FORGE, LYRA, TECTON, AXIOM, VECTOR, DATOS, QUORUM!routing_claude.md— SIR Scanner v3.3, bilingual routing!contract_builder.md— 9-Step prompt construction framework!intent_engine.md,!debug_engine.md— 36 patterns + A-P error taxonomy!templates_library.md,!writing_suite.md,!visual_suite.md!vendors_tier1-4.md— model specs (Claude/GPT · Gemini/Grok · DeepSeek/Qwen · Kimi/GLM)!memory_bridge.md,!mentor_method.md,!sandbox_user.md!global_index.md,!domain_knowledge.md,user_context.md
Architecture
Foundation → Safeguards → Execution · modular · cross-model · constraints-not-pressure
Features
/lang— toggle output language (EN ↔ RU)- ARENA A/B testing framework
- Chain Orchestrator (RESEARCH_DRAFT_REVIEW, CODE_PIPELINE, CROSS_VALIDATE)
- Error taxonomy A-P with injection remedies
- Memory Bridge for cross-session state
- Contract Mode (Claude 4.x literal execution)
Quick Start
- Create a Claude Project
- Upload all 22 files from the archive to Project Knowledge
- Start a conversation — type
startor/start - Select menu item 1-29 or type your task
Models tested
Claude Opus/Sonnet 4.6 · Gemini 3.1 Pro · GPT-5.4 · DeepSeek R1/V3.2 · Kimi K2.5 · GLM-5
P2P v1.1 · first public release · 2026-04-18