
Postcept
io.github.Postcept由 GitHub 账号 Postcept 发布v0.3.3更新于 Oct 24, 2026
Verify agent actions against the system of record and get signed completion receipts.
概览
Postcept MCP 让助手核实退款、取消订阅、工单处理等高风险操作是否真正完成,并取得签名回执。
- 可以做什么
- 核实退款、订阅取消与支持工单状态是否与记录系统一致,每项检查可选沙箱测试模式。按编号获取历史核实记录及其签名回执。对历史核实记录重新对照实时记录系统进行复核。读取组织的已验证完成率。
- 适用场景
- 适用于代理执行退款、取消订阅或工单处理等有后果的操作,且需要可核查的完成证明时,尤其适合开发、演示和已信任代理判断的工作流。
- 运行要求
- 通过 npx 以 stdio 进程在本地运行(需要 Node.js)。需要具备 verifications:write 权限的 Postcept API 密钥,通过环境变量 POSTCEPT_API_KEY 提供。可选 POSTCEPT_BASE_URL 指向自建或预发布 API;需要能访问 Postcept API 的网络。
安装
在 SourceWeft 中
- 打开 控制台中的 Postcept,将其添加到工作区。
- 为需要使用其工具的对话启用该服务。
Desktop only,通过 STDIO。 STDIO 服务会启动本地进程,因此需要 SourceWeft 桌面宿主。
其他 MCP 客户端
参照 仓库 中的启动说明。
README
@postcept/mcp
A Model Context Protocol server that exposes Postcept, outcome verification for AI agents, as MCP tools.
After an agent performs a high-risk action (a refund, a subscription
cancellation, a support-ticket resolution), it calls a verify_* tool to confirm
the action actually completed in the system of record (Stripe, Zendesk, and more)
and obtain a signed completion receipt. "done" becomes proof, not a claim.
The server is a thin wrapper over the live Postcept HTTP API. It holds no credentials of its own beyond your Postcept API key, and verification is read-only against your systems of record.
Configuration
Create an API key in the Postcept dashboard under API keys.
Use with Claude Desktop / Claude Code
Tools
Every verify_* tool accepts test: true to run against the deterministic
sandbox connector (excluded from your Verified Completion Rate), so an agent can
exercise the full flow without a live system of record.
A non-verified result (incomplete, duplicated, mismatched,
policy_failed) means the work is not done. Surface the gap and recover.
Advisory, by design
MCP puts the verification call in the agent's hands, which makes it the right integration for development, demos, and workflows where the agent's judgment is already trusted. It is advisory: an agent can skip the call, pass the wrong identifiers, or ignore the answer.
For consequential actions, enforce the decision outside the agent. The workflow
or orchestrator calls the Postcept API (or the SDK's guard()) itself and
branches on safe_to_claim_complete before anything customer-facing happens.
The agent then receives the allowed outcome instead of deciding it. Both
patterns use the same API and the same receipts, what changes is who owns the
branch.
Development
来源:README.md,提交 7089af4
工具
0版本历史
1- v0.3.3最新Sep 16, 2026
评分
已满足适用检查项的 50%
该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。
- 近 30 天内可连通25 分
请保持端点能够响应 MCP 握手;SourceWeft 每 30 天探测一次。
- 已声明认证方式10 分
请在注册中心条目中声明凭据请求头,或保持端点能够响应探测,以便判定其认证方式。
- 已列出工具15 分
请让服务器在无凭据时响应 tools/list,或发布可在沙箱中无凭据启动的软件包。
- 描述必需15 分
- 图标10 分
- 文档15 分
- 许可证10 分