AgentSkillsCN

monoco_workflow_research

Spike研究工作流（Flow Skill）。定义从添加外部仓库，到知识提取与归档的标准操作流程，确保外部知识得到有效管理与利用。

SKILL.md
--- frontmatter
name: monoco_workflow_research
description: Spike Research Workflow (Flow Skill). Defines the standard operational process from adding external repositories to knowledge extraction and archiving, ensuring effective management of external knowledge.
type: workflow
domain: spike
version: 1.0.0

Research Workflow

Standardized workflow for Spike research, ensuring the "Add → Sync → Analyze → Extract → Archive" process.

Workflow State Machine

mermaid
stateDiagram-v2
    [*] --> Add: Discover reference repository
    
    Add --> Sync: Configuration added
    Add --> Add: Invalid URL<br/>(re-enter)
    
    Sync --> Analyze: Sync completed
    Sync --> Sync: Sync failed<br/>(network/permission)
    
    Analyze --> Extract: Analysis completed
    Analyze --> Analyze: Content mismatch<br/>(re-evaluate)
    
    Extract --> Archive: Knowledge extraction completed
    
    Archive --> [*]: Archiving completed

Execution Steps

1. Add (Add Repository)

  • •Goal: Add external repository as reference
  • •Input: Repository URL, reference purpose
  • •Output: Configuration updated
  • •Checkpoints:
    • • Verify repository URL is accessible
    • • Confirm repository is relevant to current project
    • • Run monoco spike add <url>
    • • Check .monoco/config.yaml is updated

2. Sync (Synchronize)

  • •Goal: Download or update reference repository content
  • •Checkpoints:
    • • Run monoco spike sync
    • • Verify repository is cloned to .references/<name>/
    • • Check sync logs for errors
    • • Verify file permissions are correct (read-only)

3. Analyze (Analysis)

  • •Goal: Research the structure and content of reference repository
  • •Strategy: Systematic browsing and marking
  • •Checkpoints:
    • • Browse overall repository structure
    • • Identify modules relevant to current project
    • • Mark valuable code patterns
    • • Record architectural design highlights

4. Extract (Knowledge Extraction)

  • •Goal: Extract usable knowledge from reference repository
  • •Strategy: Document valuable findings
  • •Checkpoints:
    • • Extract key code snippets (without modifying original files)
    • • Record design patterns and best practices
    • • Create learning notes (using Memo or Issue)
    • • Annotate knowledge source (repository URL + Commit)

5. Archive (Archiving)

  • •Goal: Organize and archive research results
  • •Checkpoints:
    • • Update project documentation, referencing the repository
    • • Create knowledge index (if needed)
    • • Run monoco spike list to verify status
    • • Periodically clean up references no longer needed

Decision Branches

ConditionAction
Invalid URLReturn to Add, check URL format
Sync failedCheck network, permissions, retry or skip
Content not relevantRemove from configuration, reselect
Important pattern discoveredCreate Issue, plan to introduce to project

Compliance Requirements

  • •Prohibited: Editing any files in .references/
  • •Required: All external knowledge must be annotated with source
  • •Required: Regular synchronization to get updates
  • •Recommended: Only add high-quality, relevant repositories

Related Commands

bash
# Add reference repository
monoco spike add <url>

# Sync all repositories
monoco spike sync

# List configured repositories
monoco spike list

# Record research findings
monoco memo add "Discovery: {insight}" -c "spike:{repo_name}"

Best Practices

  1. •Selective Choice: Only add repositories highly relevant to the project
  2. •Read-only Access: Treat .references/ as external knowledge base, never modify
  3. •Regular Sync: Run monoco spike sync monthly to get updates
  4. •Knowledge Transformation: Transform learned patterns into actual project improvements