# Skills Library · _INDEX

The menu of every skill this brain can run. The AI reads THIS file whenever
a job might match a skill, and opens a skill file only when a trigger fires.
That is the L1/L2 deal: this index stays small and always loaded, the detail
stays in files opened on demand.

Downloaded once from https://loai.wisdomblend.com/materials/library/
The online library grows weekly. The `library-update` skill checks what is new.

## Rules for the AI

- A trigger matches -> open that skill's file and follow it exactly.
- After my weekly review: compare the week's biggest friction against this
  list and offer AT MOST ONE skill, in one sentence, with what it would have
  saved me. If nothing clearly fits, offer nothing. Never offer two.
- A skill not listed here does not exist. New skill = new line here first.
- Keep this file under about 4,000 characters and the Active table under
  20 lines. Detail lives in the skill files, never here.

## Active skills

| Skill | Trigger | What it does | Requires | Level |
|---|---|---|---|---|
| [morning-brief](morning-brief.md) | "brief me" | Calendar + email + today's priorities in one brief under 200 words | Calendar (read), Email (read) | Optimum |
| [weekly-review-focus](weekly-review-focus.md) | "run my weekly review" | Scores the week 0-10 per priority against evidence, drafts my Focus file update | Files; Calendar (read) sharpens it | Minimum |
| [gmail-weekly-email](gmail-weekly-email.md) | "email pass" | The week's inbox in 4 lists, replies drafted, nothing sent without OK | Email (read + draft) | Optimum |
| [library-update](library-update.md) | "check the library" | Diffs this index against the online library, reports new skills, offers max one | Web fetch | Minimum |
| [system-audit](system-audit.md) | "audit my system" | Full read-only audit: loaded files in tokens, skills, tools tested live, security holes, top 5 fixes | File access; strongest model, max thinking | Optimum |
| [content-audit](content-audit.md) | "audit my content" | The librarian pass: every folder has one honest index, constitution within token budget, 5 layers checked | File access; strongest model, max thinking | Minimum |
| [graduation-check](graduation-check.md) | "grade my system" | Scores all 5 layers 0-100 with evidence, 3 binary gates, verdict MINIMUM/OPTIMUM/NOT YET, pass = 75 per layer | File access; strongest model, max thinking | Minimum |

How every skill is written (the fixed format): [SKILL-FORMAT.md](SKILL-FORMAT.md)

## Coming (online library roadmap)

These are not loaded and cannot be triggered. They are the shipping plan.
Run `library-update` to pull each one when it goes live.

| Skill | One-liner | Level | Ships with |
|---|---|---|---|
| calls-analysis | A call transcript in; decisions, promises and follow-ups out | Optimum | M4 · Skill · Calls & email analysis |
| meeting-prep | Reads the invite, past emails and notes per attendee, hands you one page before the call | Optimum | M4 · Skill · Calls & email analysis |
| post-creation | One idea in, a post drafted in your voice with 3 hook options | Optimum | M4 · Skill · Post creation (planned, follows Calls & email analysis) |
| focus-email-analysis | Scores your inbox week against P1: which emails moved it, which stole from it | Optimum | M4 · Skill · Calls & email analysis |
| image-brief | Turns a post or page into a precise brief for whatever image tool you use | Optimum | M4 · Creator tools |
| site-build | From client calls to an app-builder brief: pages, copy blocks, style, done-definition | Expert | M5 · The client website machine |
| site-maintain | Weekly site check: uptime, broken links, stale pages, basic SEO, one fix list | Optimum | M5 · The client website machine |
| seo-content | Client voice + keyword list in, a batch of drafted posts out, publish only after OK | Expert | M5 · The client website machine |
| social-media-analysis | Reads your last 30 days of posts and tells you what worked, numbers not vibes | Expert | M5 · APIs & integrations |