Index of every page on this site: llms.txt
All skills
cat-xierluo avatar

/magpie-optimizer

@5fb69a0
Run it yourself

用 Magpie 配置并优化 AI 工作效能:接入供应商、订阅与插件,配置模型路由组与意图规则,管理网关、可见性、共享 MCP 与搜索;盘点额度与套餐,诊断 Token 消耗与限流,把每次调整关联真实任务验收效果。

Use this Skill: https://skilld.dev/gh/cat-xierluo/legal-skills/magpie-optimizer

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

referencesoutcomes.md

≈644 tokens on demand. Your agent reads this file only when SKILL.md points to it.

任务、成果与效果复核

使用现有任务源记录输入版本、验收条件和任务ID。观察中的session_ref/project_ref是摘要标识,不能靠标题自动猜任务归属。必要时由用户或可信任务记录明确关联;一个任务可含多个会话,同一份请求不可分摊两次。

以下命令在 Skill 目录内执行,scripts/、config/、state/ 均相对该目录。

输入

{
  "schema": "magpie-outcomes.v1",
  "records": [{
    "task_id": "task-before-1",
    "comparison_key": "same-input-class-and-acceptance",
    "stage": "before",
    "observation": "<skill-dir>/state/observation.json",
    "sessions": ["session_ref_from_observation"],
    "artifact": "<skill-dir>/state/actual-output.md",
    "acceptance": "<skill-dir>/state/acceptance.json",
    "rework_count": 0,
    "human_minutes": 10
  }]
}

验收文件:

{
  "schema": "magpie-acceptance.v1",
  "task_id": "task-before-1",
  "artifact_sha256": "hash-of-current-artifact",
  "accepted": true,
  "reviewed_by": "named-reviewer",
  "criteria": ["来源正确且关键项完整"]
}

具名评审负责内容质量;工具检查真实文件与验收哈希,不能把生产者随便写的accepted当独立语义证明。验收材料应来自已有审查/用户认可过程,不为凑对比自动签发。用户选择的comparison_key声明任务类别与验收可比性,脚本不证明任务同难度。

复核

python3 scripts/magpie_outcomes.py --manifest state/outcomes.json --output state/review.json

必须是完整分页观察且能匹配全部会话。实际文件变化、未绑定验收、重复请求分摊均拒绝。缺before/after或质量未通过输出INSUFFICIENT_EVIDENCE,不补默认“省钱”。before/after都有通过证据时,按每份验收成果比较调用数、非缓存输入、输出、缓存读取、累计请求耗时、返工、人工时间与完整已定价估算。价格缺失保持null,不把目录价当账单;累计请求耗时不等于并发任务墙钟时间。

输出DESCRIPTIVE_COMPARISON只描述观测差异,causal_improvement_proven始终为false。扩展多项任务、明确能力/质量底线,检查输入难度、人工介入、平台和模型变动等混杂因素后,才能形成有限个人建议;不自动推广成全局最佳模型。

现有经验库或任务账本保存结论及证据引用、适用任务、版本、窗口、样本量与重新评估条件。本Skill不另建竞争性任务数据库,不自动关闭业务任务。

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts18d3 checks · Risk SAFE
  • Gen Agent Trust Hub18d

    This skill provides a robust and security-conscious framework for managing AI model routing and telemetry via the Magpie CLI. It includes extensive defensive measures such as local-loopback-only network enforcement, restricted file permissions (0600), comprehensive credential redaction logic, and a self-auditing suite to prevent secret leakage.

  • Socket18d

    No alerts

  • Snyk18d

    Risk: LOW · No issues

Signed by skilld at 5fb69a0. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated yesterday
Other metadata
metadata
{
  "version": "1.8.1",
  "author": "杨卫薪律师(微信ywxlaw)",
  "homepage": "https://github.com/cat-xierluo/legal-skills"
}

README badge

README badge for cat-xierluo/legal-skills/magpie-optimizer