Token导航 LogoToken导航

Agent Skill · 研究检索

imc-plan(综合管理计划)

imc-plan 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

imc-plan 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为执行命令;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 60,976。

总安装 360周安装 15GitHub Stars 1下载 120核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:imc-plan(综合管理计划)
来源仓库:https://github.com/hungv47/marketing-skills
仓库路径:skills/imc-plan
安装命令:
npx skills add https://github.com/hungv47/marketing-skills --skill imc-plan
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/hungv47/marketing-skills --skill imc-plan

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词、任务场景或来源线索快速定位候选结果。使用时需要确认搜索范围与来源可靠性;涉及敏感或专有信息时,应先核对访问权限与合规边界。

安装方式:github,命令为 npx skills add https://github.com/hungv47/marketing-skills --skill imc-plan。建议结合具体任务关键词验证搜索结果的相关性与准确性。

来源仓库内路径:skills/imc-plan

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/hungv47/marketing-skills --skill imc-plan;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要执行命令,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

imc-plan 怎么安装?

在终端执行 npx skills add https://github.com/hungv47/marketing-skills --skill imc-plan。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

imc-plan 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

imc-plan 需要什么权限?

权限级别是执行命令。安装流程涉及命令执行,可能通过 npx skills add https://github.com/hungv47/marketing-skills --skill imc-plan 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

imc-plan 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

imc-plan 用的人多吗?

公开数据是累计安装 360 次,近一周安装 15 次,下载 120 次,GitHub 1 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

imc-plan 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要执行命令,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 不能接受 Skill 在你机器上执行命令的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
imc-plan

综合管理计划 · 本页

3601执行命令未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

IMC Plan — Orchestrator

*Communication — Step 2 of 4. Coordinates specialized agents to turn ICP research into an integrated marketing communication strategy.*

Core Question: "Does every angle trace to a pillar, every channel to a habitat, and every timeline slot to a real team capacity?"

Critical Gates — Read First

  • Identify the growth motion BEFORE selecting channels. PLG (product drives acquisition), SLG (outbound/performance drives acquisition), or Hybrid. The motion determines channel priorities — PLG favors community/SEO/forums, SLG favors paid/email/IRL/SMS.
  • Evaluate ALL 9 channels, not just digital. The full channel map: Search engines/AEO, Store/Listing platforms, Bounty/Info platforms, News, Forums/Communities, Social media, IRL (OOH/Events/POS), Mailbox, SMS. Offline channels (IRL, SMS) produce strategy docs and creative briefs for physical execution.
  • Do NOT generate angles before pillars. Angles are derived PER PILLAR. Without pillars, angles are untethered and fail the anti-generic test.
  • Do NOT assign channels without habitat data. Channel selection comes from ICP research habitat maps, not marketer preference. No habitat data → interview for it.
  • Do NOT schedule 10 pieces/week for a 2-person team. Match cadence to actual capacity. Over-scheduling guarantees missed deadlines.
  • Stale ICP research (>30 days) produces misaligned plans. Recommend re-running icp-research before proceeding.

Philosophy

Frameworks (ORB, 3D Angles, Pillar Types) are proven defaults — not mandatory templates. Adapt based on data. When ICP research provides habitat maps, use them. When it doesn't, gather the data before planning channels.

Inputs Required

  • ICP research from research/icp-research.md + research/product-context.md (or user-provided)

Output

  • .agents/mkt/imc-plan.md

Quality Gate

Before delivering, the critic agent verifies:

  • Growth motion explicitly stated (PLG / SLG / Hybrid)
  • 3-5 pillars, each with ICP evidence
  • 3+ angles per pillar, each passing 3Q test and scored ≥15/25
  • All 9 channels evaluated (selected or explicitly skipped with rationale)
  • Each selected channel has ONE specific angle (not a content category)
  • Channel selection based on ICP habitat data AND growth motion alignment
  • Channel execution briefs present for every selected channel
  • Offline channels (if selected) include compliance and execution notes
  • Timeline has 3 phases with realistic cadence
  • Launch sequence follows ORB (Owned → Rented → Borrowed)

Chain Position

Previous: icp-research | Next: content-create Re-run triggers: When ICP research is updated, when launching a new product/campaign, or when attribution reveals underperforming channels.

Skill Deference

  • Need actual content written? → Use content-create — this skill plans, not writes.
  • Need numeric targets (CAC, LTV, conversion rates)? → Use funnel-planner.
  • Need to compare pages for SEO? → Coordinate with seo for keyword targeting and content structure.
  • Need to attribute results to initiatives? → Use attribution after execution.

Agent Manifest

AgentLayerFileFocus
Pillar Agent1agents/pillar-agent.md3-5 messaging pillars from ICP pains
Angle Agent2 (sequential)agents/angle-agent.md3D angle generation per pillar
Channel Agent2 (sequential)agents/channel-agent.md9-channel evaluation with habitat-informed selection + execution briefs
Timeline Agent2 (sequential)agents/timeline-agent.mdPhase sequencing + editorial calendar
Launch Sequencing Agent2 (sequential)agents/launch-sequencing-agent.mdORB Framework channel activation order
Critic Agent2 (final)agents/critic-agent.mdAlignment, scoring rigor, completeness

Shared References (read by agents)

  • references/3d-angle-framework.md — Three-dimensional angle generation methodology (used by angle-agent)
  • references/channel-strategy.md — Habitat-informed channel selection framework (used by channel-agent)
  • references/examples.md — Complete worked examples across 5 business types

Note: This skill is primarily sequential — each agent depends on the previous. Pillar-agent is the only Layer 1 agent. The value of multi-agent here is in specialist focus, critic gate, and single-agent fallback — not parallelism.


Routing Logic

Route A: Quick Plan (MVP or Startup)

When: Limited ICP data, small team, need a plan fast.

1. Pre-dispatch: Gather context (Step 0)
2. Dispatch: pillar-agent (3 pillars, not 5)
3. Dispatch: angle-agent (2 angles per pillar, not 3+)
4. Dispatch: channel-agent (top 2-3 channels only)
5. Dispatch: critic-agent
6. If FAIL → re-dispatch (max 2 cycles)
7. Deliver — timeline and launch sequence done by orchestrator inline

Route B: Full Plan

When: ICP research complete, campaign launch, or strategic planning.

1. Pre-dispatch: Gather context (Step 0)
2. LAYER 1: Dispatch pillar-agent
3. LAYER 2 — Dispatch SEQUENTIALLY:
   - angle-agent (receives pillar output)
   - channel-agent (receives angle output + habitat data)
   - timeline-agent (receives channel output)
   - launch-sequencing-agent (receives timeline output)
4. Dispatch: critic-agent (receives complete plan)
5. If FAIL → re-dispatch named agent(s) with feedback (max 2 cycles)
6. Deliver artifact

Route C: Called by Another Skill

When: content-create or attribution needs plan context.

1. Read existing .agents/mkt/imc-plan.md if available
2. If not available, run Route B
3. Return plan to calling skill

Step 0: Pre-Dispatch Context Gathering

Product Context Check

Check for research/product-context.md and research/icp-research.md. If date fields are older than 30 days, warn the user and recommend re-running icp-research. Tip: /navigate status (from meta-skills) gives a single-pass freshness report across all upstream artifacts.

Required Artifacts

ArtifactSourceIf Missing
icp-research.mdicp-researchINTERVIEW. Gather: audience, pains, habitats, awareness stage.

Optional Artifacts

ArtifactSourceBenefit
product-context.mdicp-researchProduct details, voice adjectives
solution-design.mdsolution-designStrategic initiatives for alignment
mkt/content-research.mdcontent-research (research-skills)Competitor content patterns, audience language, trending topics — informs channel selection and angle priorities

Growth Motion Identification

Before dispatching any agent, determine the growth motion:

SignalMotionChannel Priority
Free tier/trial exists, self-serve signup, product viralityPLGSearch engines/AEO, Forums/Communities, Social media (organic), Mailbox (onboarding), Store/Listing platforms, Bounty/Info platforms
Sales team, high ACV, demos required, paid channels primarySLGSearch engines/AEO (paid), Social media (paid), Mailbox (outbound), SMS, IRL (events/OOH), News
Both motions coexist (e.g., self-serve SMB + sales enterprise)HybridBlend both — designate primary motion per audience segment

If growth motion is unclear from upstream artifacts, ask: "Does your product grow through self-serve adoption (PLG), outbound/paid performance (SLG), or both?"

Context to Pass to All Agents

  1. Growth motion: PLG, SLG, or Hybrid (determines channel weighting)
  2. Campaign goal: what success looks like
  3. ICP summary: persona, pains, habitats, awareness levels
  4. VoC quotes: top buyer-language phrases
  5. Constraints: team size, budget, timeline

Dispatch Protocol

How to spawn a sub-agent

  1. Read the agent instruction file — include its FULL content in the Agent prompt
  2. Append the context and upstream output after the instructions
  3. Resolve file paths to absolute: replace relative paths with absolute paths
  4. Pass upstream artifacts by content: orchestrator reads .agents/ files, includes relevant excerpts
  5. If feedback exists, append with "## Critic Feedback — Address Every Point"

Single-agent fallback

If multi-agent dispatch is unavailable, execute sequentially in-context:

  1. Pillar extraction (3-5 from ICP)
  2. Angle generation per pillar (3D framework)
  3. Channel assignment per angle (habitat-informed)
  4. Timeline phasing (Pre-launch → Launch → Sustain)
  5. Launch sequencing (ORB)
  6. Self-evaluate with critic rubric

Orchestrator-Written Sections

The orchestrator writes these sections inline (no agent dispatched):

  • Growth Motion — determined in Step 0, written directly into the artifact before Layer 1
  • Foundation — core message and awareness distribution, assembled after pillar-agent returns
  • Channel Execution Briefs — assembled by the orchestrator after channel-agent returns, using channel-agent's output (channel assignments + habitat data) to populate the briefs table with objective, tactic, budget type, metric, owner, and milestone per channel

Layer 1: Pillar Foundation

Dispatch pillar-agent alone. It produces the pillar table that all downstream agents need.

AgentInstruction FilePass These InputsReference Files
Pillar Agentagents/pillar-agent.mdbrief + ICP research summary + VoC—

Wait for completion. Pillar output becomes input for all Layer 2 agents.


Layer 2: Sequential Strategy Chain

Dispatch ONE AT A TIME, IN ORDER. Each receives the previous agent's output.

pillar-agent → angle-agent → channel-agent → timeline-agent → launch-sequencing-agent → critic-agent
StepAgentInstruction FileReceivesReference Files
1Angle Agentagents/angle-agent.mdPillar tablereferences/3d-angle-framework.md
2Channel Agentagents/channel-agent.mdAngle bank + habitat data + growth motionreferences/channel-strategy.md (includes 9-Channel Framework)
3Timeline Agentagents/timeline-agent.mdChannel assignments—
4Launch Sequencing Agentagents/launch-sequencing-agent.mdTimeline—
5Critic Agentagents/critic-agent.mdComplete assembled plan—

Critic Gate

  • PASS: Deliver the plan artifact.
  • FAIL: Re-dispatch the earliest failing agent with critic feedback. Max 2 cycles.

Artifact Template

---
skill: imc-plan
version: 1
date: [today's date]
status: draft
---

# IMC Plan: [Campaign / Product Name]

**Date:** [today]
**Skill:** imc-plan
**Goal:** [campaign objective]
**Audience:** [primary persona]

## Growth Motion
- **Motion:** [PLG / SLG / Hybrid]
- **Primary acquisition lever:** [product / outbound / both]
- **Channel weighting rationale:** [why these channels match the motion]

## Foundation
- **Core message:** [one sentence]
- **Awareness distribution:** [% per stage]

## Pillars
| # | Pillar | Type | % | Stage | Evidence |
|---|--------|------|---|-------|---------|
| 1 | ... | ... | ... | ... | ... |

## Angle Bank
| # | Angle | Hook | Stage | Trigger | Score | Class | Pillar |
|---|-------|------|-------|---------|-------|-------|--------|
| 1 | ... | ... | ... | ... | ... | ... | ... |

## Channel Assignments
| Channel | Type | Angle | Role | Cadence |
|---------|------|-------|------|---------|
| ... | ... | ... | ... | ... |

**9-Channel Evaluation:** For each channel, the channel-agent assesses relevance and produces a select/skip decision. Channels: Search engines/AEO, Store/Listing platforms, Bounty/Info platforms, News, Forums/Communities, Social media, IRL (OOH/Events/POS), Mailbox, SMS.

## Channel Execution Briefs

For each selected channel, a mini-brief with actionable next steps:

| Channel | Objective | Tactic | Budget Type | Success Metric | Owner | First Milestone |
|---------|-----------|--------|-------------|----------------|-------|-----------------|
| [e.g. Search engines] | [e.g. Capture intent traffic] | [e.g. Google Search ads + SEO content] | [Paid + Organic] | [CTR, CPL] | [Person] | [e.g. First campaign live by W2] |

For offline channels (IRL, SMS), include execution notes:
- **IRL:** Vendor/location requirements, lead capture method (QR code, signup form), follow-up workflow
- **SMS:** Compliance requirements (TCPA/GDPR opt-in), character limits (160 GSM-7), unsubscribe mechanism
- **OOH:** Readability specs, vanity URL/QR for tracking, legal disclaimers

## Timeline
| Week | Phase | Channel | Angle | Format | Status |
|------|-------|---------|-------|--------|--------|
| W1 | ... | ... | ... | ... | Planned |

## Launch Sequence
| Phase | Timing | Channels | Action |
|-------|--------|----------|--------|
| Internal | T-4w | ... | ... |

> On re-run: rename existing artifact to `imc-plan.v[N].md` and create new with incremented version.

Next Step

Run content-create to produce the content assets from this plan. Run attribution to set up measurement. If content-research was not run before this plan, consider running it now to validate content angles with market data before creating assets.


Worked Example — B2B SaaS Project Management Tool (Route B)

Step 0: Pre-Dispatch

Goal: Drive 500 trial signups in 60 days. Audience: Engineering managers at 50-200 person companies. Awareness: 60% Problem Aware, 30% Solution Aware, 10% Product Aware.

Layer 1: Pillar Agent

Returns 4 pillars:

#PillarType%Stage
1"The meeting tax"Problem35%Problem Aware
2"Async-first shipping"Transformation30%Solution Aware
3"Teams who switched"Trust25%Product Aware
4"The async-first movement"Social10%All

Layer 2, Step 1: Angle Agent

Returns 12 angles across 4 pillars, scored 15-23/25. Example angles:

  • "12 hrs/week lost to status theater" (Data, Problem Aware, Fear, 22/25, Shareable)
  • "How to ship 2x without hiring" (How-to, Solution Aware, Greed, 20/25, Searchable)

Layer 2, Step 2: Channel Agent

Returns assignments: LinkedIn (primary, "12 hrs/week"), Email (owned, "async-first shipping"), X/Twitter (ecosystem, "teams who switched"), Blog (owned, searchable how-to content).

Layer 2, Step 3: Timeline Agent

Returns 3 phases: Pre-launch (W1-W3, problem content), Launch (W4-W5, transformation + proof), Sustain (W6+, trust + social). Weekly calendar with pillar rotation.

Layer 2, Step 4: Launch Sequencing Agent

Returns ORB sequence: Internal (T-4w) → Email list alpha (T-2w) → Partner posts (T-1w) → LinkedIn/X launch (Day 0) → PR + paid (T+1w).

Critic Agent → PASS

All quality gates met. Pillar-angle-channel traces verified. Timeline realistic for 3-person team.


Anti-Patterns

Channel-first planning — "We need TikTok content." INSTEAD: Start with habitat maps. If ICP density on TikTok is Low, deprioritize it regardless of what's trendy.

Identical messages across channels — Same text on LinkedIn, X, and Email. INSTEAD: Each channel gets one specific angle adapted to its native format.

Angles without pillar connection — Orphan angles that don't trace back. INSTEAD: Every angle names its parent pillar. No exceptions.

Timeline without capacity check — 10 pieces/week for a solo marketer. INSTEAD: Match cadence to team size. 3 pieces/week is realistic for one person.

Too many pillars — 7 pillars dilute the message. INSTEAD: 3-5. If you can't cut, you haven't committed to a strategy.


Agent Files

Sub-Agent Instructions (agents/)

Shared References (references/)

平台分布

Codex

36.55%
按下载量换算44

Claude

30.23%
按下载量换算36

Cursor

18.73%
按下载量换算22

Gemini CLI

9.34%
按下载量换算11

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

执行命令

安装流程涉及命令执行,可能通过 npx skills add https://github.com/hungv47/marketing-skills --skill imc-plan 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills