📊 /audit: Four-Cs Diagnosis
Watch the score rise. Run /audit, understand the score, prioritize gaps by leverage, and turn the diagnosis into action every week.
Illustrative values. Your /audit generates your AIOS's actual scorecard.
Detailed content
What /audit does
O /audit is a diagnostic skill that scans your AIOS structure and returns a score from 0 to 100. It reads — never writes — every file that makes up the system: manual, memory, skills, agents, MCPs, decisions, and references. The central question is simple: "is the AIOS well-built?"
✓ What /audit does
- ✓Reads manuals, memory, skills, agents, and MCPs
- ✓Score each C from 0 to 25 (total 100)
- ✓List gaps prioritized by leverage
- ✓Save
audits/audit-{data}.md(optional)
✗ What /audit DOESN’T do
- ✗Doesn’t write any system files (read-only)
- ✗Doesn’t run skills or agents during the audit
- ✗Doesn’t evaluate the quality of the automations themselves
- ✗Doesn’t suggest new automations (that’s /level-up)
Structural, not capability
There is a critical distinction that most users confuse the first time: /audit measures whether the AIOS is well built (structure), not whether it can do something new (capability). The question "could you build an automatic daily brief?" is about /level-up, not /audit.
🎯 The Fundamental Division
- → Does CLAUDE.md exist, and does it have completed sections?
- → Are all 7 connection domains covered?
- → Is there at least 1 recurring trigger?
- → Are at least 3 skills working?
- → Should there be a weekly briefing skill?
- → Do you need an email triage agent?
- → Is CRM update automation missing?
- → What is the next highest-impact automation?
If /audit returns gaps like "you could build X," note them and take them to /level-up. The audit is a mirror for structure—not a strategy consultant. Using them in the same session creates scope confusion.
How the 4 Cs are scored
Each C is scored from 0 to 25. The /audit checks specific criteria in each dimension—it’s not subjective. Below are the exact criteria that determine your score.
Stages — score ranges
The total score (0–100) falls into 4 stages. Each stage has a name, a range, and a practical interpretation of the AIOS state. Most new projects fall between 40 and 70—that’s expected progress, not a failure.
| Score | Stage | Name |
|---|---|---|
| 0–39 | Stage 0 | Foundation |
| 40–69 | Stage 1 | Built |
| 70–89 | Stage 2 | Compounding |
| 90–100 | Stage 3 | Autonomous |
Gaps by leverage
/audit doesn’t list gaps in random order. It uses a leverage formula to prioritize what to address first: leverage = points lost × impact multiplier. The 3 highest-leverage gaps come with a concrete next step.
Example: prioritized Top-3 gaps
Email and calendar aren’t connected. They block higher-ROI automations.
adicionar MCP Gmail + Google CalendarIdentity and decision sections are nearly empty. Every session starts from scratch.
expandir seções Business, Voice, DecisionsOnly the 3 skills in the kit. None customized for the business.
/level-up → criar 1 skill de brief semanalThe gap list may have 8–12 items. Focus only on the Top-3 in leverage. Fixing the first usually raises the score by 10–15 points and changes the ranking of the others. Re-run /audit after each improvement.
Read-only + scoreboard
/audit is strictly read-only—it never changes system files. The only optional write effect is saving the result to audits/audit-{data}.md to track score changes week by week. The format uses bars ## every 5 points with descriptive labels.
✓ Scoreboard best practices
- ✓Save the audit every week in
audits/ - ✓Compare the current score with last week’s
- ✓Celebrate every +5 points — that’s real progress
✗ Common pitfalls
- ✗Don’t ask /audit to “go easy”—honesty is the point
- ✗Don’t try to resolve 8 gaps at once
- ✗Don’t forget to rerun after each improvement
Weekly ritual
/audit is the compounding hook of AIOS. The first run establishes the baseline on Day 7. From then on, you run it every week, resolve the Top-1 leverage item, and watch the score climb. This is the cycle that turns a setup into a system.
Weekly audit cycle
Run /audit for the first time. Accept the score without judging it. Save it in audits/audit-semana-1.md. This is ground zero.
Open the latest audit.md. What was the Top-1 leverage item? Was it resolved? Note the status before opening a new session.
Run the skill. A new score is generated. Compare it with last week's. How much did it increase? A new Top-3 leverage list is identified.
Take the Top-1 leverage point from /audit and move it straight to /level-up. One gap → one automation. Close the loop for the week.
Resolve 1 high-leverage gap per week. In 4 weeks, an AIOS Stage 1 (57 pts) typically reaches Stage 2 (75+ pts). The rising score is evidence that the system is working — not the individual automation.
📊 Module Summary
Next Module: 3.3 — /level-up
You understand the diagnosis. Now you’ll learn to build 1 automation per week using /level-up — the growth engine of your AIOS. Each run = 1 artifact. Each week = a more capable AIOS.