Progmune Documentation 开发者文档
English below ｜ 中文见下English
Progmune is an AI Trust Decision Engine for AI-generated software. Install with npm install progmune-runtime and verify a file with npm run sdk src/server.ts --explain.
- README — what it detects, the two paths (intercept / verify), full usage.
- Quick Start — 5-minute onboarding.
- API Reference — SDK verify / explain / getCompatibility, MCP tools, npm bins.
- Coverage Matrix — language & framework coverage with evidence links.
- Runtime Architecture — trust engine internals, scoring model, decision thresholds.
- Benchmarks — protocol blind-test methodology and baselines.
- Audit Response — response to three independent third-party audits (2026-09).
MCP server: expose Progmune to Claude Code via progmune_check and progmune_trust_check tools — see the README MCP section for setup.
中文
Progmune 是面向 AI 生成软件的 AI 信任决策引擎。安装：npm install progmune-runtime；验证单个文件：npm run sdk src/server.ts --explain。
- README 中文版 — 检测内容、两条路径（拦截 / 验证）、完整用法。
- 快速开始 — 5 分钟上手。
- API 参考 — SDK verify / explain / getCompatibility、MCP 工具、npm bin。
- 覆盖矩阵 — 语言与框架覆盖现状（附证据链接）。
- 运行时架构 — 信任引擎内部、评分模型、决策阈值。
- 基准 — 协议盲测方法学与基线。
- 审计回应 — 三份独立第三方审计（2026-09）的回应。
MCP server：通过 progmune_check 与 progmune_trust_check 工具把 Progmune 接入 Claude Code——配置方式见 README 的 MCP 章节。