Agent Skills › vectorize-io/hindsight › hindsight-recall

hindsight-recall

GitHub

通过 Hindsight MCP 工具检索长期记忆，辅助用户查询过往编码决策、项目上下文及偏好。

hindsight-integrations/cursor/skills/hindsight-recall/SKILL.md vectorize-io/hindsight

触发场景

询问过往决策查询项目上下文查找历史偏好

安装

npx skills add vectorize-io/hindsight --skill hindsight-recall -g -y

更多选项

非标准路径

npx skills add https://github.com/vectorize-io/hindsight/tree/main/hindsight-integrations/cursor/skills/hindsight-recall -g -y

不安装直接使用

npx skills use vectorize-io/hindsight@hindsight-recall

指定 Agent (Claude Code)

npx skills add vectorize-io/hindsight --skill hindsight-recall -a claude-code -g -y

安装 repo 全部 skill

npx skills add vectorize-io/hindsight --all -g -y

预览 repo 内 skill

npx skills add vectorize-io/hindsight --list

SKILL.md

Frontmatter

{
    "name": "hindsight-recall",
    "description": "Search long-term memory for relevant context from past coding sessions using Hindsight MCP tools"
}

Hindsight Recall

Trigger

Use when the user explicitly asks about past decisions, project context, preferences, or anything that may have been discussed in prior sessions.

Workflow

Identify the key topic or question from the user's request.
Check if <hindsight_memories> in the current context already contains relevant results from the automatic session-start recall.
If session memory already covered it, use those memories directly.
If the user needs more specific or deeper recall, use the Hindsight MCP recall tool to search for additional memories.
If you want to reason over accumulated memories for architectural decisions, use the reflect tool.

Guardrails

Only use MCP tools for deeper recall when the session-start memory was insufficient.
When memories conflict with current context, prefer current context and note the discrepancy.
Do not expose raw memory metadata to the user unless asked.