TailorAU/pact

PACT is an extremely early-stage project (37 days old, 1 star, 0 forks, no recent velocity) that proposes an open standard for multi-agent collaboration. While the problem space is legitimate—coordinating multiple AI agents on shared documents is an emerging need—the project shows zero adoption signals and no evidence of production implementation or technical differentiation. The README positions this as a protocol/standard rather than a working system, which means it's more aspirational specification than defensible implementation. Frontier labs (especially OpenAI with GPT-4V, Anthropic, and Google with Gemini) are actively exploring agentic collaboration frameworks and multi-turn reasoning. They have orders of magnitude more resources to standardize protocols and could trivially integrate consensus mechanisms into platform features. The project lacks: (1) working reference implementation evidence, (2) empirical validation of the protocol, (3) community adoption or collaboration, (4) technical novelty (consensus mechanisms are well-studied). Without demonstrable traction, a shipping implementation, or a unique technical approach, this reads as a standards proposal document rather than a defensible technical project. Even if the protocol has merit, frontier labs would more likely subsume this into their own agent orchestration layers than adopt an external standard. Risk of obsolescence is high because agent coordination will likely become a platform feature rather than a standalone tool. Recommend: wait for concrete implementation, empirical results, and early adopter feedback before re-scoring.