Install any skill in seconds. Free to start, no credit card required.
Get Started Free →hacklyc/myagents-memory-update
仅当系统或用户明确指定完整名称 `myagents-memory-update` 时使用; 不要根据任务语义或相似表述自行触发。
Passed 5 of 11 test cases with the skill, and 7 of 11 without it.
and 1 more in the full report
3 fixed6 passed either way5 still failing8 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 3 test cases the skill fixed. Same model, same prompt, with and without the skill.
During a general chat session, the user mentioned a bug fix in authentication headers, but the current context has no explicit link to the existing mobile-app topic workspace. Should this memory update be placed into the mobile-app topic file?
| Checked | Without skill | With skill |
|---|---|---|
| Keep the entry in Daily memory | Missed | Passed |
| Do not write the update into the mobile-app topic file without explicit context link | Passed | Passed |
The instructions your model reads when the skill loads.
在当前 Session 内完成一次增量记忆巩固。目标不是总结聊天或尽量多记,而是让未来 Session 通过工作区接续这里造成的真实状态变化和高置信度协作经验。
执行前先锁定最终输出:完成所有工具调用后,只能发送调用方规定的成功标记,不发送任何 过程、检查或状态说明。
只回顾当前 Session 自上次成功更新后新增的工作。先读取工作区记忆规则、相关现有记忆 和实际产物,再回答五个问题:
错误的长期记忆通常比暂时缺失更有害。区分用户明确表达、可观察的互动事实和模型推断, 并用完整的前后文判断反馈:
用户明确提出的长期要求,或跨情境反复出现且理由一致的模式,才适合进入稳定的自动加载 记忆。证据不足时,在 Daily/Topic 中保留互动事实与不确定性,或不写;不要把推断升级成 用户画像。记录有适用情境的未来行为,而不是人格、心理或动机标签。
test cases passed, vs 7 / 11 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.