AOF Self-Evaluation Dashboard
What this is: A self-measurement system for the Agent Operating Framework (AOF) — a set of rules, hooks, and behavioral guidelines that govern how Claude Code operates. This dashboard tracks whether each rule is earning its keep, using a quality-engineering method borrowed from manufacturing.
How to read this dashboard
DPMO = Defects Per Million Opportunities (a Six Sigma quality measurement)
Six Sigma is a manufacturing quality discipline with 30+ years of industrial use. Its core idea: measure defect rates per unit of exposure, not raw counts. Applied here — a “defect” is a rule violation; an “opportunity” is every moment that rule could have fired. DPMO = (violations ÷ opportunities) × 1,000,000. This makes rates comparable across rules that fire at very different frequencies.
The signal is the rate, not the count
A rule that checked 855 times and blocked 3 has a different health profile than one that checked 5 times and blocked 3. Raw counts mislead — DPMO corrects for volume. Counts below 30 opportunities are marked ⚠ and suppressed: not enough data yet to compute a meaningful rate.
★ Load-bearing rare rules are a special class
Some rules fire rarely but guard against catastrophic outcomes (e.g., blocking a commit that would expose credentials). These are tagged ★ and excluded from “retire by low DPMO” logic — a low fire rate is expected and correct.
Disclosure: Rate-normalized signal, not causal proof — windows may be confounded by concurrent rule changes, sample size, or work-type mix.
Live signal — last 7 days
332 total rule checks recorded across 27 active rules. Top by volume: Plan Mode gate (36 checks) · Hindsight recall (31 checks) · Skill router gate (31 checks).
Load-bearing (rare fire OK)
| Rule | Type | 7d opportunities | 30d opportunities | 60d opportunities |
|---|---|---|---|---|
| PTU-1Secret pattern gate★ | hook | n=7 ⚠ | n=19 ⚠ | 78 |
| PTU-15smokin-ops migration gate★ | hook | n=0 ⚠ | n=1 ⚠ | n=1 ⚠ |
| PTU-16ST migration gate★ | hook | n=0 ⚠ | n=1 ⚠ | n=1 ⚠ |
| PTU-18Independent review gate (H6)★ | hook | n=0 ⚠ | n=0 ⚠ | n=0 ⚠ |
| PTU-3settings.xml guard★ | hook | n=3 ⚠ | n=7 ⚠ | 32 |
| PTU-4Source-of-truth write gate★ | hook | n=28 ⚠ | 51 | 110 |
Standard
| Rule | Type | 7d opportunities | 30d opportunities | 60d opportunities |
|---|---|---|---|---|
| PTU-10No-guess gate + tier format | hook | n=28 ⚠ | 51 | 110 |
| PTU-11No-guess breadcrumb | hook | n=28 ⚠ | 66 | 245 |
| PTU-12Delivery checklist gate | hook | n=0 ⚠ | n=0 ⚠ | n=0 ⚠ |
| PTU-19Claim-evidence gate (Gate 4) | hook | n=25 ⚠ | 122 | 529 |
| PTU-2Honest-word gate (T2 ban, dispatch) | hook | n=8 ⚠ | n=20 ⚠ | 79 |
| PTU-5Obsolete scoring column gate | hook | n=28 ⚠ | 51 | 110 |
| PTU-6Exchange-First breadcrumb | hook | n=0 ⚠ | n=10 ⚠ | n=10 ⚠ |
| PTU-7Exchange-First gate | hook | n=0 ⚠ | n=0 ⚠ | n=0 ⚠ |
| PTU-8Plan Mode gate | hook | 36 | 135 | 483 |
| PTU-9CC safety net | hook | n=16 ⚠ | 102 | 450 |
| SS-1Hindsight session open | hook | n=7 ⚠ | 59 | 144 |
| SS-2MCP health check | hook | n=7 ⚠ | 61 | 146 |
| SS-3Gate 6 ADR check-in reminder | hook | n=7 ⚠ | 60 | 145 |
| SS-4AGENTS.md spike reminder | hook | n=7 ⚠ | 60 | 144 |
| SS-5Workspace ID setter | hook | n=7 ⚠ | 60 | 144 |
| SS-6Pipeline-F health check | hook | n=7 ⚠ | 60 | 145 |
| SS-7Skill catalog freshness check | hook | n=7 ⚠ | 60 | 145 |
| SS-8TP daily feedback poll | hook | n=7 ⚠ | 60 | 144 |
| SS-9Communication rules inject | hook | n=7 ⚠ | 59 | 144 |
| UPS-1Hindsight recall | hook | 31 | 176 | 632 |
| UPS-2Skill router gate | hook | 31 | 177 | 633 |
Opportunity distribution — 30d
Each bar = one rule. Length = how many times it checked in the last 30 days. Amber = load-bearing rare (★). Faded = below the 30-check threshold.
30-day opportunity counts per rule. ★ = load-bearing-rare (Goodhart guard — low count is expected and OK). ⚠ = n < 30 (below threshold; percentage suppressed).
Sensitivity windows (total opportunities across all rules)
▸Rule adherence trend (all time)
▸Session cost (all time)
Cost data not populated yet. Handoff frontmatter needs cost_usd field; cc-analytics will fill this on future sessions.
▸Sessions
| Date | Machine | Rule | Plan→Delivery | Cost $ | Composite |
|---|---|---|---|---|---|
| 2026-06-24 | mac | 9.0 | — | — | 9.0 |
| 2026-06-24 | mac | 9.0 | — | — | 9.0 |
| 2026-06-20 | win | 10.0 | — | — | 10.0 |
| 2026-06-13 | mac | 10.0 | — | — | 10.0 |
| 2026-06-13 | mac | 8.5 | — | — | 8.5 |
| 2026-06-13 | mac | 9.0 | — | — | 9.0 |
| 2026-06-13 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-06-13 | mac | 9.0 | — | — | 9.0 |
| 2026-06-13 | mac | 6.0 | — | — | 6.0 |
| 2026-06-13 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-06-13 | mac | 10.0 | — | — | 10.0 |
| 2026-06-12 | mac | 9.0 | — | — | 9.0 |
| 2026-06-11 | mac | 10.0 | — | — | 10.0 |
| 2026-06-10 | mac | 9.0 | — | — | 9.0 |
| 2026-06-10 | mac | 10.0 | — | — | 10.0 |
| 2026-06-10 | mac | 10.0 | — | — | 10.0 |
| 2026-06-09 | mac | 10.0 | — | — | 10.0 |
| 2026-06-09 | mac | 6.0 | — | — | 6.0 |
| 2026-06-09 | mac | 9.0 | — | — | 9.0 |
| 2026-06-09 | mac | 6.0 | — | — | 6.0 |
| 2026-06-09 | mac | 9.0 | — | — | 9.0 |
| 2026-06-09 | mac | 10.0 | — | — | 10.0 |
| 2026-06-09 | mac | 10.0 | — | — | 10.0 |
| 2026-06-09 | mac | 10.0 | — | — | 10.0 |
| 2026-06-08 | mac | 9.0 | — | — | 9.0 |
| 2026-06-08 | mac | 9.0 | — | — | 9.0 |
| 2026-06-08 | mac | 10.0 | — | — | 10.0 |
| 2026-06-08 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 10.0 | — | — | 10.0 |
| 2026-06-07 | mac | 10.0 | — | — | 10.0 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 7.5 | — | — | 7.5 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 6.0 | — | — | 6.0 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 7.5 | — | — | 7.5 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 10.0 | — | — | 10.0 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 9.0 | — | — | 9.0 |
| 2026-06-07 | mac | 7.5 | — | — | 7.5 |
| 2026-06-07 | mac | 7.5 | — | — | 7.5 |
| 2026-06-07 | win | 9.0 | — | — | 9.0 |
| 2026-06-07 | win | 7.5 | — | — | 7.5 |
| 2026-06-07 | win | 9.0 | — | — | 9.0 |
| 2026-06-07 | win | 9.0 | — | — | 9.0 |
| 2026-06-06 | mac | 10.0 | — | — | 10.0 |
| 2026-06-06 | mac | 9.0 | — | — | 9.0 |
| 2026-06-06 | mac | 9.0 | — | — | 9.0 |
| 2026-06-06 | mac | 9.0 | — | — | 9.0 |
| 2026-06-06 | mac | 9.0 | — | — | 9.0 |
| 2026-06-06 | win | 10.0 | — | — | 10.0 |
| 2026-06-05 | mac | 10.0 | — | — | 10.0 |
| 2026-06-05 | mac | 9.0 | — | — | 9.0 |
| 2026-06-05 | mac | 9.0 | — | — | 9.0 |
| 2026-06-05 | mac | 9.0 | — | — | 9.0 |
| 2026-06-05 | mac | 9.0 | — | — | 9.0 |
| 2026-06-05 | win | 7.5 | — | — | 7.5 |
| 2026-06-05 | win | 10.0 | — | — | 10.0 |
| 2026-06-05 | win | 9.0 | — | — | 9.0 |
| 2026-06-05 | win | 10.0 | — | — | 10.0 |
| 2026-06-05 | win | 9.0 | — | — | 9.0 |
| 2026-06-05 | win | 9.0 | — | — | 9.0 |
| 2026-06-05 | win | 10.0 | — | — | 10.0 |
| 2026-06-04 | mac | 7.5 | — | — | 7.5 |
| 2026-06-04 | mac | 7.5 | — | — | 7.5 |
| 2026-06-04 | mac | 9.0 | — | — | 9.0 |
| 2026-06-04 | win | 7.5 | — | — | 7.5 |
| 2026-06-04 | win | 9.0 | — | — | 9.0 |
| 2026-06-02 | mac | 10.0 | — | — | 10.0 |
| 2026-06-02 | mac | 10.0 | — | — | 10.0 |
| 2026-05-31 | mac | 10.0 | — | — | 10.0 |
| 2026-05-31 | win | 9.0 | — | — | 9.0 |
| 2026-05-29 | win | 9.0 | — | — | 9.0 |
| 2026-05-29 | win | 10.0 | — | — | 10.0 |
| 2026-05-29 | win | 6.0 | — | — | 6.0 |
| 2026-05-29 | win | 8.5 | 10.0 | — | 9.1 |
| 2026-05-29 | win | 7.5 | — | — | 7.5 |
| 2026-05-28 | win | 9.0 | — | — | 9.0 |
| 2026-05-25 | win | 8.5 | 10.0 | — | 9.1 |
| 2026-05-25 | win | 10.0 | — | — | 10.0 |
| 2026-05-25 | win | 9.0 | — | — | 9.0 |
| 2026-05-24 | win | 7.0 | — | — | 7.0 |
| 2026-05-24 | win | 4.0 | — | — | 4.0 |
| 2026-05-24 | win | 9.0 | — | — | 9.0 |
| 2026-05-24 | win | 7.0 | — | — | 7.0 |
| 2026-05-24 | win | 4.0 | — | — | 4.0 |
| 2026-05-24 | win | 10.0 | 10.0 | — | 10.0 |
| 2026-05-23 | win | 9.0 | — | — | 9.0 |
| 2026-05-23 | mac | 9.0 | — | — | 9.0 |
| 2026-05-22 | mac | 10.0 | — | — | 10.0 |
| 2026-05-22 | win | 7.5 | — | — | 7.5 |
| 2026-05-22 | win | 10.0 | — | — | 10.0 |
| 2026-05-22 | win | 9.0 | — | — | 9.0 |
| 2026-05-22 | win | 9.0 | 10.0 | — | 9.4 |
| 2026-05-22 | win | 8.5 | 10.0 | — | 9.1 |
| 2026-05-22 | win | 9.0 | — | — | 9.0 |
| 2026-05-22 | win | 10.0 | 10.0 | — | 10.0 |
| 2026-05-22 | win | 8.5 | — | — | 8.5 |
| 2026-05-22 | win | 9.0 | — | — | 9.0 |
| 2026-05-22 | win | 7.5 | — | — | 7.5 |
| 2026-05-22 | win | 10.0 | — | — | 10.0 |
| 2026-05-22 | win | 10.0 | — | — | 10.0 |
| 2026-05-22 | win | 9.0 | — | — | 9.0 |
| 2026-05-22 | win | 9.0 | — | — | 9.0 |
| 2026-05-22 | win | 10.0 | — | — | 10.0 |
| 2026-05-21 | win | 10.0 | — | — | 10.0 |
| 2026-05-21 | win | 9.0 | — | — | 9.0 |
| 2026-05-21 | win | 9.0 | 10.0 | — | 9.4 |
| 2026-05-21 | win | 10.0 | — | — | 10.0 |
| 2026-05-21 | win | 8.5 | — | — | 8.5 |
| 2026-05-21 | win | 7.5 | — | — | 7.5 |
| 2026-05-21 | win | 10.0 | — | — | 10.0 |
| 2026-05-20 | win | 10.0 | — | — | 10.0 |
| 2026-05-20 | win | 6.0 | — | — | 6.0 |
| 2026-05-19 | win | 10.0 | 10.0 | — | 10.0 |
| 2026-05-19 | win | 10.0 | — | — | 10.0 |
| 2026-05-19 | win | 7.5 | — | — | 7.5 |
| 2026-05-19 | win | 7.0 | — | — | 7.0 |
| 2026-05-18 | win | 7.5 | — | — | 7.5 |
| 2026-05-18 | win | 7.0 | — | — | 7.0 |
| 2026-05-18 | win | 8.5 | — | — | 8.5 |
| 2026-05-18 | win | 9.0 | — | — | 9.0 |
| 2026-05-18 | win | 9.0 | — | — | 9.0 |
| 2026-05-18 | win | 9.0 | — | — | 9.0 |
| 2026-05-17 | win | 9.0 | — | — | 9.0 |
| 2026-05-17 | win | 9.0 | — | — | 9.0 |
| 2026-05-17 | win | 9.0 | — | — | 9.0 |
| 2026-05-17 | win | 10.0 | — | — | 10.0 |
| 2026-05-17 | win | 10.0 | — | — | 10.0 |
| 2026-05-16 | win | 9.0 | — | — | 9.0 |
| 2026-05-16 | win | 10.0 | — | — | 10.0 |
| 2026-05-15 | mac | 10.0 | — | — | 10.0 |
| 2026-05-15 | mac | 10.0 | — | — | 10.0 |
| 2026-05-15 | mac | 10.0 | 10.0 | — | 10.0 |
| 2026-05-15 | mac | 10.0 | — | — | 10.0 |
| 2026-05-15 | win | 8.5 | — | — | 8.5 |
| 2026-05-15 | win | 10.0 | — | — | 10.0 |
| 2026-05-15 | win | 9.0 | 10.0 | — | 9.4 |
| 2026-05-15 | win | 10.0 | — | — | 10.0 |
| 2026-05-14 | mac | 10.0 | — | — | 10.0 |
| 2026-05-14 | win | 9.0 | — | — | 9.0 |
| 2026-05-13 | mac | 9.0 | — | — | 9.0 |
| 2026-05-13 | mac | 6.0 | 10.0 | — | 7.6 |
| 2026-05-13 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-05-13 | win | 10.0 | — | — | 10.0 |
| 2026-05-12 | mac | 10.0 | 10.0 | — | 10.0 |
| 2026-05-12 | mac | 5.5 | 8.8 | — | 6.8 |
| 2026-05-12 | mac | 10.0 | — | — | 10.0 |
| 2026-05-12 | mac | 8.5 | — | — | 8.5 |
| 2026-05-11 | mac | 10.0 | — | — | 10.0 |
| 2026-05-11 | mac | 10.0 | — | — | 10.0 |
| 2026-05-11 | mac | 10.0 | — | — | 10.0 |
| 2026-05-11 | mac | 10.0 | — | — | 10.0 |
| 2026-05-11 | mac | 8.5 | 10.0 | — | 9.1 |
| 2026-05-11 | mac | 7.0 | 7.5 | — | 7.2 |
| 2026-05-11 | mac | 10.0 | — | — | 10.0 |
| 2026-05-11 | mac | 10.0 | — | — | 10.0 |
| 2026-05-11 | mac | 10.0 | — | — | 10.0 |
| 2026-05-11 | mac | 10.0 | — | — | 10.0 |
| 2026-05-10 | mac | 7.5 | — | — | 7.5 |
| 2026-05-10 | mac | 7.5 | — | — | 7.5 |
| 2026-05-10 | win | 9.0 | — | — | 9.0 |
| 2026-05-09 | mac | 9.0 | — | — | 9.0 |
| 2026-05-09 | mac | 9.0 | — | — | 9.0 |
| 2026-05-09 | mac | 9.0 | — | — | 9.0 |
| 2026-05-09 | mac | 9.0 | — | — | 9.0 |
| 2026-05-08 | mac | 9.0 | — | — | 9.0 |
| 2026-05-08 | mac | 9.0 | — | — | 9.0 |
| 2026-05-08 | mac | 9.0 | — | — | 9.0 |
| 2026-05-08 | mac | 10.0 | — | — | 10.0 |
| 2026-05-08 | mac | 9.0 | — | — | 9.0 |
| 2026-05-08 | mac | 9.0 | — | — | 9.0 |
| 2026-05-08 | mac | 9.0 | — | — | 9.0 |
| 2026-05-08 | mac | 10.0 | — | — | 10.0 |
| 2026-05-07 | mac | 10.0 | — | — | 10.0 |
| 2026-05-07 | mac | 7.5 | — | — | 7.5 |
| 2026-05-07 | mac | 10.0 | — | — | 10.0 |
| 2026-05-07 | mac | 9.0 | — | — | 9.0 |
| 2026-05-07 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-05-07 | mac | 10.0 | — | — | 10.0 |
| 2026-05-07 | mac | 7.5 | — | — | 7.5 |
| 2026-05-07 | mac | 10.0 | — | — | 10.0 |
| 2026-05-07 | mac | 9.0 | — | — | 9.0 |
| 2026-05-07 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-05-06 | mac | 9.0 | — | — | 9.0 |
| 2026-05-06 | mac | 9.0 | — | — | 9.0 |
| 2026-05-05 | mac | 10.0 | — | — | 10.0 |
| 2026-05-05 | mac | 10.0 | 10.0 | — | 10.0 |
| 2026-05-05 | mac | 10.0 | — | — | 10.0 |
| 2026-05-05 | mac | 10.0 | 10.0 | — | 10.0 |
| 2026-05-04 | mac | 9.0 | — | — | 9.0 |
| 2026-05-04 | mac | 7.5 | — | — | 7.5 |
| 2026-05-04 | mac | 9.0 | — | — | 9.0 |
| 2026-05-04 | mac | 7.5 | — | — | 7.5 |
| 2026-05-03 | mac | 10.0 | — | — | 10.0 |
| 2026-05-03 | mac | 9.0 | — | — | 9.0 |
| 2026-05-03 | mac | 8.5 | — | — | 8.5 |
| 2026-05-03 | mac | 10.0 | — | — | 10.0 |
| 2026-05-03 | mac | 10.0 | — | — | 10.0 |
| 2026-05-03 | mac | 9.0 | — | — | 9.0 |
| 2026-05-03 | mac | 8.5 | — | — | 8.5 |
| 2026-05-03 | mac | 10.0 | — | — | 10.0 |
| 2026-05-02 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-05-02 | mac | 10.0 | — | — | 10.0 |
| 2026-05-02 | mac | 8.5 | — | — | 8.5 |
| 2026-05-02 | mac | 10.0 | — | — | 10.0 |
| 2026-05-02 | mac | 6.0 | — | — | 6.0 |
| 2026-05-02 | mac | 10.0 | — | — | 10.0 |
| 2026-05-02 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-05-02 | mac | 10.0 | — | — | 10.0 |
| 2026-05-02 | mac | 8.5 | — | — | 8.5 |
| 2026-05-02 | mac | 10.0 | — | — | 10.0 |
| 2026-05-02 | mac | 6.0 | — | — | 6.0 |
| 2026-05-02 | mac | 10.0 | — | — | 10.0 |
| 2026-05-01 | mac | 10.0 | — | — | 10.0 |
| 2026-05-01 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-05-01 | mac | 10.0 | — | — | 10.0 |
| 2026-05-01 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-04-30 | mac | 8.5 | — | — | 8.5 |
| 2026-04-30 | mac | 9.0 | — | — | 9.0 |
| 2026-04-30 | mac | 9.0 | — | — | 9.0 |
| 2026-04-29 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-04-29 | mac | 10.0 | — | — | 10.0 |
| 2026-04-29 | mac | 9.0 | — | — | 9.0 |
| 2026-04-28 | mac | 9.0 | — | — | 9.0 |
| 2026-04-28 | mac | 7.5 | — | — | 7.5 |
| 2026-04-28 | mac | 9.0 | 10.0 | — | 9.4 |
| 2026-04-28 | mac | 10.0 | — | — | 10.0 |
| 2026-04-28 | mac | 10.0 | — | — | 10.0 |
| 2026-04-27 | mac | 9.0 | — | — | 9.0 |
| 2026-04-27 | mac | 9.0 | — | — | 9.0 |
| 2026-04-27 | mac | 9.0 | — | — | 9.0 |
Current vs baseline — latest harness run 2026-06-28 (2026-05-19 → 2026-06-28, manual)
The most recent graded run compared against the frozen v1.5 baseline (May 24, 2026). Green = above baseline, amber = below. Both use the same v1.0 session-composite grader, so the deltas are like-for-like.
Historical baseline — v1.5 baseline · last harness run May 24, 2026
Scored by the original harness (v1.0), which measured rule adherence, plan→delivery gap, and cost per session. Frozen as the pre-DPMO reference point.
This baseline predates the DPMO framework. The two measurement systems are complementary: the composite score above measured session-level quality (0–10); the DPMO panels above measure rule-level utilization and violation rate. Both are required to understand whether the framework is working.