AI Agent Hub
Back to skills
Workflow Memory Hub icon

Workflow Memory Hub

AI Agent Updated 2026.08.29

Paste the following prompt into your AI chat to install this skill:

Follow https://skillhub.cn/install/skillhub.md to install @user_6e3bb2d0/workflow-memory.

About this skill

Problem

Recurring tasks often depend on one-off instructions, leaving tool choices, parameters, and step order scattered across chat history. WorkflowHub externalizes these repeated working patterns into a local workflow library, with files as the source of truth instead of burying process knowledge in a temporary prompt.

How It Works And Boundaries

The skill operates in four stages: match the current request, confirm reuse, compile an execution brief, and save or update the workflow. Matching uses keywords, toolchain overlap, and recent session actions. Before execution, it generates a brief with the goal, ordered steps, and preferred tools rather than pasting raw JSON.

Key capabilities include:
- match_workflows.py for deterministic similarity checks
- render_workflow_prompt.py for drafting briefs
- new_workflow.py, capture_workflow.py, and save_workflow.py for capturing and saving
- ui/server.py for a local editable UI

Boundaries: it is not intended for one-off micro tasks. A match only suggests reuse; it does not auto-execute. Workflows with writes, external calls, or other side effects must be explicitly requested in the current turn.

Use Cases

  • After a code review, save the checklist, tool order, and recurring keywords as a reusable `pr-review` workflow.
  • When a similar review request arrives, match the saved workflow and ask whether to follow that SOP.
  • While producing repeated data reports, draft the actual steps into a JSON workflow with `capture_workflow.py`.
  • In the local UI, search and edit workflow keywords, steps, tool preferences, and last-updated metadata.

Best For

  • Engineers who want to save recurring code-review processes as local SOPs
  • Analysts or operations staff who repeat similar reports and want reusable tool and parameter order
  • Developers using AI agents for multi-step tasks and wanting a local editor for workflow files
  • Agent users who want personal cross-project work habits stored as diff-friendly JSON