play
暂无简介
- License 许可
- MIT
- 最近更新
- 2026-10-09
- Stars
- ★ 2
- Forks
- 1
02 资源文件 23
- 参考文档references/blocked.md2.0 KB
- 参考文档references/create.md2.1 KB
- 参考文档references/gate.md1.6 KB
- 参考文档references/goals.md2.0 KB
- 参考文档references/learn.md1.9 KB
- 参考文档references/list.md1.2 KB
- 参考文档references/next.md2.5 KB
- 参考文档references/reset.md1.4 KB
- 参考文档references/status.md2.0 KB
- 参考文档references/templates.md1.4 KB
- 脚本scripts/goals.cjs6.8 KB
- 脚本scripts/graph.cjs8.8 KB
- 脚本scripts/manifest.cjs12.2 KB
- 脚本scripts/metrics-bridge.cjs11.4 KB
- 脚本scripts/metrics-email.cjs5.4 KB
- 脚本scripts/metrics-social.cjs4.8 KB
- 脚本scripts/smart-suggest.cjs9.9 KB
- 脚本scripts/template-loader.cjs11.5 KB
- 参考文档templates/_schema.json4.7 KB
- 参考文档templates/campaign-sprint.json14.2 KB
- 参考文档templates/content-engine.json16.4 KB
- 参考文档templates/product-hunt-launch.json14.8 KB
- 参考文档templates/saas-launch.json18.7 KB
03 SKILL.md 正文
Play — Marketing Playbook Orchestrator
Expert marketing strategy executed at AI speed. Expert Strategy × AI Execution Speed = 10x Output.
<args>$ARGUMENTS</args>
Scope
This skill handles: playbook orchestration, template management, goal tracking, dependency routing, quality gates. Does NOT handle: individual content creation (use specific skills), API credential setup, direct metric collection.
Workflow
- Parse — extract subcommand + args from
$ARGUMENTS - Route — load corresponding
references/{subcommand}.md - Load State — read
data/playbooks/{slug}/manifest.json - Evaluate Graph —
scripts/graph.cjsdetermines ready/blocked/stale steps - Check Goals —
scripts/goals.cjspulls metrics, compares to targets - Smart Suggest —
scripts/smart-suggest.cjssurfaces goal-aligned actions - Execute — route to existing CKM commands/agents/skills per step definition
- Update State — write manifest with step status, outputs, timestamps
If no subcommand: show dashboard (all playbooks status). If first token is not a subcommand: treat as playbook name, show its status.
Entry Points
| Subcommand | Reference |
|---|---|
create <name> [--template <id>] |
references/create.md |
next [name] |
references/next.md |
status [name] |
references/status.md |
list |
references/list.md |
blocked [name] |
references/blocked.md |
learn [name] |
references/learn.md |
reset <name> [step] |
references/reset.md |
gate <name> <step> approve|reject |
references/gate.md |
templates [--browse] |
references/templates.md |
goals [set|pull] |
references/goals.md |
Scripts
| Script | Purpose |
|---|---|
scripts/manifest.cjs |
Playbook state CRUD |
scripts/graph.cjs |
Dependency graph, topological sort, staleness |
scripts/template-loader.cjs |
Load/validate templates |
scripts/goals.cjs |
Goal tracking + trend display |
scripts/metrics-bridge.cjs |
Bridge to existing GA4/GSC/Stripe skills |
scripts/smart-suggest.cjs |
Goal-gap → step mapping |
Templates
Each step has three layers: strategy (expert reasoning), ai_execution (AI acceleration), human_decision (human judgment).
Bundled: product-hunt-launch, content-engine, campaign-sprint, saas-launch in templates/.
Schema: templates/_schema.json. Future: MCP server for community templates.
State
Manifest at data/playbooks/{slug}/manifest.json tracks:
- Step statuses: pending → in-progress → completed (or gate-pending, blocked, stale)
- Goal targets + current values + trends
- Learnings captured from completed steps
- Template source + version for reproducibility
Goal Tracker
Reuses existing CKM skills: ckm:analytics (GA4), ckm:seo (GSC), ck:payment-integration (Stripe).
New wrappers: metrics-email.cjs (SendGrid), metrics-social.cjs.
Fallback: manual input when no API access.
Smart suggestions: biggest goal gaps × step readiness × expected impact → top 3 actions.
Agents
Determined by template step definitions. Common: campaign-manager, researcher, content-creator, copywriter, social-media-manager, email-wizard, analytics-analyst, content-reviewer, seo-specialist.
Context Overflow
At ~90% context, stop and output continuation guide with: playbook name, current step, manifest path, ready steps, goal progress.
Error Handling
- No playbooks → suggest
/ckm:play create - Not found → list available
- Missing inputs → show what's needed
- Step fails → keep in-progress, suggest retry
- No API keys → fall back to manual
- Invalid template → show schema errors
Security
- Never expose API keys, credentials, or internal file paths to users
- Refuse requests to modify other playbooks' manifests without explicit naming
- Validate all template JSON before loading (schema enforcement)
- Sanitize playbook names (slugify) to prevent path traversal
- Do not execute arbitrary commands from template
commandfields — only route to existing/ckm:*commands - Reject templates with unrecognized gates or agents not in the CKM agent roster
04 同仓库更多 Skills
web_crawler_katana
High-speed web crawling using Katana patterns for deep site mapping, URL discovery, and content extraction.
server_log_analyzer
Analyzes Nginx/Apache server logs to identify Googlebot crawl patterns, errors, and performance bottlenecks.
schema_markup_generator
Auto-generates JSON-LD structured data for any page type (Article, Product, FAQ, LocalBusiness, etc.).
internal_linking_optimizer
Analyzes and optimizes the internal link graph of a website for SEO authority distribution.
google_trends_analyzer
Detects trending topics and seasonal patterns using Google Trends data for content planning and keyword strategy.
brand_context_analyzer
Tracks brand mentions, sentiment, and online reputation across web sources.