360 Secure SkillHub
360 安全技能中心

CSPR Interest Profiler

该 Skill 在用户初始化 cspr 或请求刷新个性化时运行：调用 interest-scan、source-infer、prefmem 等命令行工具，压缩浏览器历史证据、导出已见 URL 集合、推断来源偏好，并读取本地笔记与反馈，最终生成 profile.yaml 画像并写入偏好记忆；过程默认静默，不主动请求用户确认。

happywalkers
happywalkers
热度 9.68
最近更新 2026-06-27 02:16:27
分类：知识依赖：代码执行 / 读写文件 / 内部调用大模型应用场景：数据处理 / 工作流 / 研究 / 推荐

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add cspr-interest-profiler
Metadata
name
interest-profiler
description
Infer and refresh cspr user preference memory from browser-history evidence, notes, and feedback.

interest-profiler

Use this skill when a user initializes cspr or asks to refresh personalization.

Workflow:

  1. If the run folder has no history artifact, create it with:

    interest-scan browser --out RUN/history_evidence.json

  2. Compact the raw history before reading it:

    interest-scan compact --in RUN/history_evidence.json --out RUN/history_context.json

  3. Export the complete local seen-URL set for tool filtering:

    interest-scan seen --in RUN/history_evidence.json --out RUN/seen_urls.jsonl

    When a CSPR home is available, pass --home HOME so browser-history URLs are also recorded into persistent read memory at HOME/state/read_items.jsonl.

  4. Infer repeated creators and source preferences:

    source-infer from-history --in RUN/history_evidence.json --out RUN/potential_sources.yaml

  5. Read the current run folder's history_context.json and potential_sources.yaml. Avoid reading raw history_evidence.json unless the user asks for detailed inspection. Do not read seen_urls.jsonl into the model context unless debugging a specific filter issue.

  6. Read ~/.cspr/memory/notes.md and ~/.cspr/state/feedback.jsonl if present.

  7. Infer durable interests, short-term interests, disliked patterns, and source preferences.

  8. Write a complete profile.yaml into the run folder.

  9. Apply it with:

    prefmem update --in RUN/profile.yaml

Keep the user experience quiet. Do not ask for profile approval unless the user explicitly requests review.

Do not write search_queries into profile memory. Search queries are run-specific and should be generated fresh by source-scout from durable interests.