brain-health

Check which brain files are complete, incomplete, or empty, and get recommendations for what to fill next. Trigger with /brain-health or when the user mentions 'brain status,' 'brain health,' 'what's

SKILL.md

GitHub

Brain Health Check

Scan all brain files, report completeness, and tell the user exactly which skill to run next to fill gaps.


Workflow

Step 1: Scan brain files

Read every file in brain/ and count [FILL IN] and [VERIFY] placeholders in each. Also check for files that exist in the expected set but are missing from the directory.

Expected brain files:

File Category
truth.md Core
positioning-and-messaging.md Core
personas.md Core
competitive.md Core
capabilities.md Core
use-cases.md Core
voice-and-tone.md Brand
voice-samples.md Brand
brand-guide/brand-guide.md Brand
audience-language.md Supporting
customer-journey.md Supporting
market-signals.md Supporting
qualifying-questions.md Supporting
tactical-assets.md Supporting

Step 2: Calculate completeness

For each file, determine status:

  • Complete — zero [FILL IN] and [VERIFY] placeholders remaining
  • Partial — some placeholders filled, some remaining. Show the count of remaining placeholders.
  • Empty — all or nearly all fields are still [FILL IN] (untouched template). Treat a file as Empty if it has no substantive content beyond the template boilerplate.
  • Missing — file doesn't exist in brain/

Step 3: Present the dashboard

Display a health report. Use a simple text-based progress bar for each file.

Brain Health Report
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Core Files
  truth.md                     ██████████  Complete
  positioning-and-messaging.md ████░░░░░░  18 fields remaining
  competitive.md               ██████░░░░  6 fields remaining
  personas.md                  ████░░░░░░  9 fields remaining
  capabilities.md              ░░░░░░░░░░  Empty
  use-cases.md                 ░░░░░░░░░░  Empty

Brand
  voice-and-tone.md            ░░░░░░░░░░  Empty — run /tone-mapping
  voice-samples.md             ░░░░░░░░░░  Empty — run /tone-mapping
  brand-guide.md               ████████░░  3 fields remaining

Supporting Files
  audience-language.md         ░░░░░░░░░░  Empty
  customer-journey.md          ░░░░░░░░░░  Empty
  market-signals.md            ░░░░░░░░░░  Empty
  qualifying-questions.md      ░░░░░░░░░░  Empty
  tactical-assets.md           ░░░░░░░░░░  Empty

Overall: 1 of 14 files complete

Step 4: Recommend next actions

After the dashboard, recommend the top 3 files to fill next, in priority order. Use this priority logic:

  1. Core files first, supporting files later. Core files feed every skill; supporting files feed specialized skills.
  2. Within core, prioritize by downstream impact:
    • truth.md (everything depends on this)
    • positioning-and-messaging.md (messaging pillars feed capabilities, use-cases, content)
    • competitive.md (feeds competitor pages, battlecards)
    • personas.md (feeds content targeting)
    • capabilities.md (feeds feature pages, datasheets)
    • use-cases.md (feeds case studies, landing pages)
  3. Brand files when core is >60% done.
  4. Supporting files when core is >80% done.

For each recommendation, specify the skill to run:

Brain File Skill to Run What to Prepare
truth.md Manual entry or /setup Company facts, product details, customer list
positioning-and-messaging.md /product-marketing Knowledge of market problem, competitors, differentiation
competitive.md /product-marketing Competitor names, why you win/lose against each
personas.md /product-marketing Buyer titles, their problems, what they care about
capabilities.md /product-marketing Product features organized by messaging pillar
use-cases.md /product-marketing Customer scenarios, triggers, outcomes
voice-and-tone.md /tone-mapping 3-5 real writing samples (LinkedIn posts, blog posts, emails)
voice-samples.md /tone-mapping (Step 6, after the profile) 5+ real writing samples — more samples, better calibration
brand-guide.md /design-extract Your website URL, or brand colors/fonts manually
audience-language.md /product-marketing Sales call transcripts or customer conversations
customer-journey.md /product-marketing Sales process knowledge, common objections, deal patterns
market-signals.md /product-marketing Industry trends, analyst reports, news
qualifying-questions.md /product-marketing Discovery call experience, qualification criteria
tactical-assets.md /product-marketing Company bios, speaker bios, social profiles

Present the recommendations:

Recommended next steps (in order):

1. capabilities.md — Run /product-marketing and ask to build out capabilities
   Have ready: your product's features organized by messaging pillar

2. voice-and-tone.md — Run /tone-mapping with 3-5 writing samples
   Have ready: LinkedIn posts, blog posts, or emails you wrote yourself

3. use-cases.md — Run /product-marketing and ask to define use cases
   Have ready: customer scenarios and the triggers that bring them to you

Step 5: Offer to start

Ask:

Want to start filling one of these now? Pick a number, or tell me which file you'd like to work on.

If they pick one, route to the appropriate skill.


Rules

  • Skip sources.md. It's a log kept by /brain-ingest, not a file to fill. If it has open items under "To check," mention the count after the recommendations as a likely next source.
  • Read files as-is. Don't modify any brain files during a health check.
  • Count [FILL IN] and [VERIFY] literally (case-insensitive match on [fill in] and [verify]).
  • If a file has custom content mixed with some remaining placeholders, it's partial, not empty.
  • Don't judge quality of filled-in content. The health check measures completeness, not correctness.
  • If running at the end of /setup, skip Step 5 (the offer to start) since setup is wrapping up. Just show the dashboard and recommendations.