A Standardized Prompt Syntax/Language to train models and humans against for faster behavior tuning of models.
HEALTH ? 78 / Stable
A Standardized Prompt Syntax/Language to train models and humans against for faster behavior tuning of models.
// readme
Repo: https://github.com/noizu-labs-ml/NoizuPromptLingo
The NPL structured prompt-engineering syntax β placeholders, intuition pumps, directives, prefixes, and tagged prompt sections β plus a public MCP that serves that corpus. This repo publishes the language agents load; the agent work platform (sessions, tickets, chat, wiki, memory, tobor MCP fleet) lives separately in agent-kit-mcp.
conventions/ β single source of truth for the NPL syntax sections (syntax, declarations, pumps, directives, prefixes, prompt-sections, special-sections)./mcp β two tools only: NPLLoad (expression DSL) and NPLSpec (versioned spec generation). No authentication required.sub=site:npl). The token does not identify a user; it only means βapproved for this site.β/api/v1/npl/* (public) for browsing sections and generating the same spec the admin panel produces.