$npx -y skills add MoizIbnYousaf/marketing-cli --skill deepen-planEnhance an existing marketing plan with parallel research. Use when the agent has a draft plan (brand strategy, campaign brief, content calendar, launch plan) that needs strengthening with real data. Triggers on 'deepen this plan', 'strengthen this strategy', 'research gaps in my
| 1 | # /deepen-plan — Plan Enhancement with Parallel Research |
| 2 | |
| 3 | The second pass. Takes an existing marketing plan and strengthens it with targeted research — audience data, competitive intel, keyword gaps — then merges findings back in. |
| 4 | |
| 5 | For brand memory protocol, see /cmo [rules/brand-memory.md](../cmo/rules/brand-memory.md). |
| 6 | |
| 7 | ## On Activation |
| 8 | |
| 9 | 1. Read `brand/` files if they exist (voice-profile.md, audience.md, competitors.md, positioning.md, keyword-plan.md). They enhance but never gate. |
| 10 | 2. Identify the plan to deepen — the user will either point to a file or have one in context. |
| 11 | |
| 12 | ## Phase 0: Assess the Plan |
| 13 | |
| 14 | Read the existing plan and classify it: |
| 15 | |
| 16 | | Plan Type | Signals | Common Gaps | |
| 17 | |-----------|---------|-------------| |
| 18 | | Brand strategy | Voice, positioning, values | Missing audience data, no competitive differentiation | |
| 19 | | Campaign brief | Timeline, channels, messages | Vague audience, no keyword backing, weak positioning | |
| 20 | | Content calendar | Topics, dates, channels | No keyword research, missing distribution strategy | |
| 21 | | Launch plan | Timeline, milestones, channels | No competitive landscape, missing pricing context | |
| 22 | | Brainstorm output | Approaches, recommendations | Everything — brainstorms are directional, not researched | |
| 23 | |
| 24 | If no plan is found or provided: |
| 25 | "I need an existing plan to deepen. Either point me to a file or run `/brainstorm` or `/launch-strategy` first to create one." |
| 26 | |
| 27 | If the plan is embedded in conversation context (not a file): extract the key sections and work with them inline. Note in the handoff that the deepened plan should be saved to a file. |
| 28 | |
| 29 | ## Phase 1: Gap Analysis |
| 30 | |
| 31 | Score the plan across 5 dimensions. For each, rate as **strong**, **weak**, or **missing**: |
| 32 | |
| 33 | ``` |
| 34 | ### Plan Gap Analysis |
| 35 | |
| 36 | | Dimension | Status | Evidence | |
| 37 | |-----------|--------|----------| |
| 38 | | Audience clarity | [strong/weak/missing] | [What the plan says or doesn't] | |
| 39 | | Competitive positioning | [strong/weak/missing] | [What the plan says or doesn't] | |
| 40 | | Keyword/search strategy | [strong/weak/missing] | [What the plan says or doesn't] | |
| 41 | | Distribution plan | [strong/weak/missing] | [What the plan says or doesn't] | |
| 42 | | Success metrics | [strong/weak/missing] | [What the plan says or doesn't] | |
| 43 | ``` |
| 44 | |
| 45 | Show this table to the user. Explain which gaps you'll fill and why they matter: |
| 46 | "Your campaign brief has strong messaging but no audience data backing it. The competitive section is a single sentence. I'm going to research both in parallel — this will take a few minutes but will make the plan 3x more actionable." |
| 47 | |
| 48 | If all dimensions are **strong**: "This plan is solid. I don't see gaps worth researching. If you want me to dig deeper on a specific section, tell me which one." |
| 49 | |
| 50 | ### Quality Gate |
| 51 | |
| 52 | A dimension is **strong** when it contains: |
| 53 | - Specific data points (numbers, names, quotes) rather than vague statements |
| 54 | - Actionable details an agent or human could execute on |
| 55 | - Evidence of research, not just assumptions |
| 56 | |
| 57 | A dimension is **weak** when it: |
| 58 | - States the obvious without specifics ("target tech-savvy users") |
| 59 | - Has the right structure but lacks data backing |
| 60 | - Makes claims without evidence |
| 61 | |
| 62 | A dimension is **missing** when: |
| 63 | - The section doesn't exist at all |
| 64 | - It's a single vague sentence |
| 65 | |
| 66 | ## Phase 2: Parallel Research |
| 67 | |
| 68 | Spawn research agents based on the gaps found. Use the Agent tool to run up to 3 agents **simultaneously in a single message**: |
| 69 | |
| 70 | ### When audience is weak or missing: |
| 71 | Spawn `mktg-audience-researcher`: |
| 72 | - Provide: the product/project name, what it does, who the plan currently targets (if stated) |
| 73 | - Agent writes to `brand/audience.md` |
| 74 | - Tell it: "Research the target audience for [product]. The current plan says [quote relevant section]. Find buyer personas, where they hang out online, what language they use, and what objections they have. Write findings to brand/audience.md." |
| 75 | |
| 76 | ### When competitive positioning is weak or missing: |
| 77 | Spawn `mktg-competitive-scanner`: |
| 78 | - Provide: the product/project name, market space, any competitors mentioned in the plan |
| 79 | - Agent writes to `brand/competitors.md` |
| 80 | - Tell it: "Research competitors for [product] in the |