技能 人工智能 目标通过评分记忆循环求解

目标通过评分记忆循环求解

v20260701
accint-solve
本技能指导用户通过AI的评分记忆循环(scored-memory loop)来解决复杂目标。它提供了一个结构化的工作流,用于处理中间的“大脑框架”(brain_frame)结果,进行系统性的推理和持续的提案提交。适用于需要高度结构化、可追溯、多步骤决策的复杂问题解决场景。
获取技能
265 次下载
概览

solve

When to Use

Use this skill when you need route a goal through acc's scored-memory loop via acc_act(runtime="solve"); deliberate any returned brain_frame and submit via continue.

Routing sugar over the two MCP verbs — no logic lives here.

  1. Call acc_act(runtime="solve", input="<the goal>").
  2. If the result is final: surface the answer, the commitment id, and the cited [ids].
  3. If the result is a brain_frame: it is YOUR deliberation turn — the frame is typed (which hole, what was retrieved, what is predicted). Reason over it, then submit via acc_act(runtime="continue", input={"frame_id": ..., "submit_token": ..., "proposal_text": ...}).
  4. End proposal_text with PREDICT: <0.00-1.00> <why>; acc strips that line before the owner sees it and uses it to calibrate the Work Model against later outcomes.
  5. Never leave a received frame unresolved; never solo-derive outside the loop.
  6. Close the commitment honestly later with acc_act(runtime="outcome", ...).

Limitations

  • Use this skill only when the task clearly matches its upstream source and local project context.
  • Verify commands, generated code, dependencies, credentials, and external service behavior before applying changes.
  • Do not treat examples as a substitute for environment-specific tests, security review, or user approval for destructive or costly actions.
信息
Category 人工智能
Name accint-solve
版本 v20260701
大小 1.83KB
更新时间 2026-07-02
语言