Skill Concept Brief — Production Workflow

How to Use This Skill

Follow this workflow for every new skill. The concept brief must be validated before opening the Skill Build Kit. Do not skip steps. Do not build the skill file first and brief after.

Step 0: Determine Mode

Before starting, determine whether this is a Handraiser or Intensive skill.

• Handraiser: Free skill for LinkedIn → DM campaign. Standalone, pasted input, works without foundations.
• Intensive: Paid skill built during a 3-session Intensive. Practice Brain input, chains with other skills, works within a session arc.

If Intensive: Read the Intensive Concept Brief first (content-pipeline/concept-briefs/intensive-concept-brief-[topic].md). It contains the arc, IP direction per session, skill definitions, and the Cohesion Check showing handraiser → Intensive upgrade paths. This is your primary source — the Intensive Concept Brief has already done the high-level IP sourcing. Your job is to go deeper on the specific skill.

Step 1: Gather IP Direction

Handraiser Mode

Ask Kathryn. The IP direction comes from her — concept titles, angles, framings. These are the primary source material.

Intensive Mode

Read the Intensive Concept Brief. The IP direction is already documented in the IP Direction table for this skill's session. Start there, then go deeper into the vault sources listed.

What to capture (both modes):

• Concept titles / working names for what the skill should do
• The problem (what practice owners don't know that they should)
• Any specific angles or framings
• What the user walks away with

Intensive mode — also capture:

• Session position in the arc (what comes before, what comes after)
• Practice Brain files the skill reads from
• Upstream skill output this skill can consume (Sessions 2-3)
• Downstream skill input this skill must produce (Sessions 1-2)
• Handraiser version of this skill, if one exists (from Cohesion Check)

Do NOT:

• Go hunting for existing angle files to replace Kathryn's direction
• Propose a direction before she's given hers
• Assume the existing IP is the right source

Step 2: Source Existing IP

After Kathryn provides direction, search for adjacent IP that informs the design. Start with the vault — that's where the deep methodology lives.

Search locations (in priority order):

1. advisory-os-vault/content/business/marketing/ip-library/ — GPT assistant system prompts (.txt files). These contain full methodology: conversation flows, diagnostic frameworks, classification systems, scoring logic, output structures. Read every .txt file that relates to the skill's problem space.
2. advisory-os-vault/content/business/marketing/ip-library/micromagnet-archive-2026-3-15/ — Micro-magnet source files (.docx). Kathryn's concept titles as fully developed guides. Binary files — note the title and that it exists even if you can't read the content.
3. advisory-os-vault/content/business/marketing/campaigns/ — 12+ campaign folders with articles, micro-tools, calculators, skill files, narratives, email sequences. IP is baked into built assets — articles explain frameworks, micro-tools contain scoring/diagnostic logic, skill files contain production methodology. Search every campaign folder for assets related to the skill's problem space.
4. advisory-os-vault/content/business/marketing/content-pipeline/concept-briefs/ — Concept briefs for micro-magnets. Check for overlapping or adjacent concepts. Verify differentiation.
5. advisory-os-vault/content/frameworks/ — existing production kits (not IP source, but may affect design)
6. business-aos/reference/proof/angles/ — named frameworks (thin reference only — vault has the real depth)
7. business-aos/reference/proof/case-studies/ — engagement arcs
8. business-aos/reference/core/ip-inventory.md — full IP catalog (cross-reference)

Search order matters. The vault IP library (1-4) contains built methodology. business-aos reference files (6-8) contain thin summaries. If the same concept appears in both, the vault version is the source.

Classify what you find:

• Source material = Kathryn's concepts (always primary)
• Deep methodology = vault GPT assistants and campaign assets with full frameworks, scoring logic, conversation flows
• Reference material = existing IP that informs but doesn't override (business-aos angle files, case studies)
• Not relevant = existing IP that looks related but serves a different purpose
• Gap = IP that should exist but doesn't. Flag it.

If there are gaps or upgrades needed: The brief must identify them explicitly.

Gap = IP that doesn't exist yet. A content interview with Kathryn is required to create it. The interview extracts the methodology from her head — frameworks, signal patterns, client examples, language she uses.

Upgrade = IP that exists but is below Kathryn's current level of thinking. Her knowledge, experience, and pattern recognition evolve continuously. What she wrote 6 months ago may be thin compared to what she knows now after more client work and testing. A content interview enriches the existing IP to match her current level.

Flag in the brief as:

IP Gap: [What's missing]. Content interview required to extract [specific methodology/framework/examples].

IP Upgrade: [What exists] at [location]. Current state: [what's thin/outdated]. Content interview required to enrich [what specifically needs deepening].

Step 3: Define the Problem

Write the problem in 3 bullets or fewer. The problem describes what the user doesn't know — not what the skill does.

Test: Read the problem statement. Would a practice owner say "that's me"? If not, rewrite.

Step 4: Map IP to Skill Design

Build the IP Direction table — every concept must trace to its vault source:

Then map signal types / output sections back to IP concepts with methodology depth:

Every section must trace. If it doesn't connect to an IP concept, question whether it belongs. If the methodology is a Gap, flag it for content interview.

Step 5: Check Design Constraints

Run each constraint:

Can't fail:

• Define the input path (what do they paste?)
• Run the zero-friction test (already have it? under 2 minutes? works messy?)
• Define a second input path if possible (Skill #1 output?)
• If any answer is no → redesign the input before proceeding

Sustainable:

• Define the repeatable cadence (weekly? monthly? quarterly?)
• What changes between runs that makes running it again valuable?
• Does the skill have a comparison section for repeat runs?

Win fast:

• What's the specific "how did I not see that" moment?
• Is the win actionable (they can do something today) or just informational?
• Does the output include ready-to-use language/scripts/actions?

Intensive Mode — Additional Constraints

Non-technical:

• Can someone who downloaded Claude yesterday install and run this?
• Where's the friction? How does the Field Guide or live session reduce it?

Practice Brain integration:

• What reference files does this skill read from?
• What happens if they're incomplete? (Must degrade gracefully — manual input fallback)
• What's the difference in output quality between Practice Brain vs. manual input?

Chains with other skills:

• What does this skill receive from the upstream skill? (Sessions 2-3)
• What does this skill produce for the downstream skill? (Sessions 1-2)
• Document the exact output sections that map to the next skill's input.

Session-scoped:

• Can they install, customize, and get output in ~50 min?
• What can be pre-configured vs. what must they customize?

If any constraint fails, stop and redesign. Do not proceed with a brief that fails a constraint.

Step 6: Check Quality Bar

Answer these questions:

1. Would someone feel fortunate they got this for free?
2. Would they feel slightly guilty they didn't pay?
3. What would this output cost from a consultant? (Must be $500+)
4. How does this pair with other skills in the series?

If the answers aren't convincing, the skill isn't ambitious enough.

Step 7: Note Foundational Dependencies

Handraiser Mode

State explicitly:

• Does this skill work WITHOUT the foundational skills (Service List, ICP, Voice)?
• What makes it BETTER with them?
• For the free campaign: it MUST work standalone
• The upgrade path: "Build the foundations in Practice Builders OS and this skill gets 10x better"

Intensive Mode

State explicitly:

• What Practice Brain files does this skill require? (client roster, services catalog, etc.)
• What happens without them? (Must still run — manual input fallback)
• What's the output quality difference? (Practice Brain vs. manual — this motivates Groundwork completion)
• Does the handraiser version of this skill exist? If yes, what's the upgrade? (Same methodology, different input modality — structured data vs. pasted emails)

Step 8: Design the Output

List output sections in a table:

Rules:

• Each section has ONE job
• ~6-10 sections (enough depth, not bloated)
• Signal types (if applicable) get their own table with IP tracing
• The output format should be scannable — tables, checklists, label/value pairs

Step 9: Check Series Cohesion

Handraiser Mode

Build the arc table:

Verify:

• Throughline is consistent across all skills
• This skill connects to the previous one (ideally: output of Skill N feeds Skill N+1)
• Series position matches what was teased in prior DMs/emails

Intensive Mode

Build the arc table from the Intensive Concept Brief:

Verify:

• Each skill's output is structured so the next skill can read it
• The chain is documented: which output sections map to which input
• The arc narrative holds (from the Intensive Concept Brief)
• Handraiser → Intensive upgrade path is documented (same job, different capability)

Step 10: Flag What's Missing

Teaching story: Mark as TBD. List the specific questions Kathryn needs to answer from testing:

• How many [signals/items/findings] did it find?
• What was the most surprising finding?
• What was the "it was right" moment?
• Did the [scripts/actions/language] feel usable?

Distribution: Trigger word, URLs, Cloudinary — all TBD at this stage.

Open questions: Number them. Be specific. These are decisions Kathryn needs to make.

Step 11: QC and Present

Run 04-skill-concept-brief-quality.md against the brief. Fix all failures.

Add the preview link to the top: Preview: https://markdownlivepreview.com/

Save to: content/business/marketing/content-pipeline/concept-briefs/skill-[##]-[name].md

Present to Kathryn for validation. Nothing gets built until she says go.