import-sop（进口标准作业程序）

/dm:import-sop

Purpose

Import and structure agency Standard Operating Procedures (SOPs) that apply across all clients. SOPs define how work gets done — approval workflows, content review steps, campaign launch checklists, escalation procedures, and quality gates.

SOPs are stored at the agency level (~/.claude-marketing/sops/), not per-brand, so they apply consistently across all clients.

Input Required

The user provides:

• SOP content: Pasted workflow steps, checklist items, or process descriptions
• SOP name: What this SOP covers (e.g., "content-approval", "campaign-launch", "crisis-escalation")
• Description (optional): Brief summary of when this SOP applies

If the user doesn't provide a name, infer it from the content.

Process

1. Load brand context: Read ~/.claude-marketing/brands/_active-brand.json for the active slug, then load ~/.claude-marketing/brands/{slug}/profile.json. Apply brand voice, compliance rules for target markets (skills/context-engine/compliance-rules.md), and industry context. Also check for existing guidelines at ~/.claude-marketing/brands/{slug}/guidelines/_manifest.json — if present, load restrictions and relevant category files. Check for custom templates at ~/.claude-marketing/brands/{slug}/templates/. Check for agency SOPs at ~/.claude-marketing/sops/. If no brand exists, ask: "Set up a brand first (/dm:brand-setup)?" — or proceed with defaults.
2. Classify the SOP type:

- Content workflow: Review, approval, and publishing steps for content - Campaign checklist: Pre-launch, launch, and post-launch verification steps - Escalation procedure: Who to contact, decision authority, response timelines - Quality assurance: Testing steps, brand compliance checks, accessibility verification - Client onboarding: Steps for setting up a new client in the system - Reporting workflow: Data collection, analysis, presentation, delivery steps

1. Structure into actionable steps:

- Number each step clearly - Identify decision points (if/then branches) - Mark which steps require human approval vs. automated checks - Add role assignments where mentioned (who does what) - Include timelines/SLAs where provided - Flag any steps the plugin can automate vs. steps requiring human action

1. Check for existing SOPs — If an SOP with this name already exists:

- Show the current SOP content - Ask: merge (combine steps), replace (overwrite), or cancel - When merging, deduplicate and maintain step order

1. Save the SOP:

- Save using guidelines-manager.py --action save-sop --name {name} --content "{content}" - Or write directly to ~/.claude-marketing/sops/{name}.md - Update the SOP manifest

1. Explain integration — Tell the user how this SOP will be applied:

- Which commands will reference this SOP - Which workflow steps will be added to outputs - When human approval gates will be flagged

Output

• Confirmation with SOP name and step count
• Preview of the structured SOP
• Explanation of which commands/modules will reference this SOP
• Suggestion: "This SOP will apply across all brands. To create brand-specific workflows, use guidelines instead."

Examples

User: "Before publishing any content, it needs to go through: 1. Writer creates draft, 2. Editor reviews for quality, 3. Brand manager checks voice alignment, 4. Legal reviews if it contains claims, 5. Client approves, 6. Publish"

Result: Saves to ~/.claude-marketing/sops/content-approval.md:

# Content Approval Workflow

## Scope
Applies to all content before publishing across all brands/clients.

## Steps

1. **Writer creates draft**
   - Role: Content Creator
   - Plugin support: Content Engine generates draft with brand voice applied
   - Output: Draft content document

2. **Editor reviews for quality**
   - Role: Editor
   - Checklist: Grammar, clarity, structure, readability score
   - Plugin support: Content scorer provides readability metrics
   - Gate: Human approval required

3. **Brand manager checks voice alignment**
   - Role: Brand Manager
   - Checklist: Voice consistency, messaging alignment, restriction compliance
   - Plugin support: Brand voice scorer provides alignment score
   - Gate: Human approval required

4. **Legal reviews (conditional)**
   - Trigger: Content contains health claims, financial claims, testimonials, or competitor comparisons
   - Role: Legal Team
   - Gate: Human approval required

5. **Client approves**
   - Role: Client Stakeholder
   - Gate: Human approval required

6. **Publish**
   - Role: Content Creator / Social Media Manager
   - Plugin support: Platform formatting applied automatically

User: "Our agency has a crisis escalation process: minor issues go to account manager, major issues go to agency director, critical issues go to CEO within 1 hour"

Result: Saves to ~/.claude-marketing/sops/crisis-escalation.md:

# Crisis Escalation Procedure

## Severity Levels

### Minor (Severity: Low)
- **Examples**: Negative review, minor social media complaint, factual error in published content
- **Escalate to**: Account Manager
- **Response SLA**: 4 hours
- **Action**: Acknowledge, draft response, resolve

### Major (Severity: Medium)
- **Examples**: Viral negative post, media inquiry, product recall mention, multiple complaints
- **Escalate to**: Agency Director
- **Response SLA**: 1 hour
- **Action**: Pause scheduled content, draft holding statement, convene response team

### Critical (Severity: High)
- **Examples**: Legal threat, data breach, executive controversy, widespread media coverage
- **Escalate to**: CEO
- **Response SLA**: 1 hour
- **Action**: All content paused, crisis team activated, holding statement within 30 minutes

Reference Files

• skills/context-engine/guidelines-framework.md — How SOPs integrate with the guideline system
• scripts/guidelines-manager.py — CLI for SOP CRUD operations (--action list-sops, save-sop, get-sop)