{{TARGET NAME}}
the 4R audit: read, run, reality, report

{{1 LINE: what was audited, and that nothing in the target was changed}}
Pulled fresh{{DATE}} · {{TIME}}
{{THE 1 NUMBER THAT MATTERS MOST}} {{WHAT THAT NUMBER COUNTS}}

{{THE VERDICT, 1 SENTENCE}}{{THE 1 FIX THAT MATTERS MOST, 1 LINE}}

{{HOW THE AUDIT RAN: the 4 layers in 2 or 3 plain sentences, the job the stranger was given, and that everything ran in a throwaway copy}}

Routes that resolve{{N of M}}{{note}}
Planted shapes caught{{N of M}}{{note}}
Stranger's job{{done / partly / no}}{{N stuck points}}
Claims still true{{N of M}}{{note}}
Work at risk{{N}}{{unsaved folders, keys in files}}
Fixes found{{N}}{{Money N, Lost work N, Slows the agent N, Polish N. Open from last time N}}

The most important

What the audit saw

3 different claims, kept apart. {{N}} files were IN the copy, each with a fingerprint on the snapshot list. {{N}} were LEFT OUT, each with its reason. {{N}} were EXAMINED: opened by the stranger or the auditor (the reader's own account). Fingerprint check: {{the VERIFY line}}. Helpers ran on: {{script runner model}}, {{stranger model}}, {{test runner model}}.

Left outSizeReasonDoes the map name this file, and did the reader go looking for this file?
{{path}}{{KB}}{{over the size limit / the ignore rule that matched}}{{no / NEEDED: named in the map at line N, brought along or NOT IN THE COPY / WANTED: what the reader was trying to do}}
{{every file left out for size, 1 row each, and 1 row per ignore rule}}

Every fix, measured and ranked ({{N}} in all, no cap)

Every finding from every layer is here. Heaviest weight first: Money, then Lost work, then Slows the agent, then Polish. Inside a weight, the bigger counted number first.

#FixWho it hitsHow much, countedWeightFile and line
1{{FIX}}{{customers and money / your work / the agent}}{{the number and what was counted}}{{Money / Lost work / Slows the agent / Polish}}{{file}}:{{line}}
{{N}}{{every other finding, 1 row each, until none is left}}

Say go and every fix gets made in this order, each with its proof, then the audit runs again. {{the calls only you can make, each as 1 plain question}}

Layer 1, read: what the files say

{{1 line per question that found something, each with its evidence}}

Layer 2, run: a stranger did 1 real job

{{the job, what came out, how many files the stranger read before knowing where to go, each stuck point in the stranger's own words with file and line}}

Layer 3, reality: planted mistakes and claims that say done

PromiseGuardOn the mistakeOn the clean caseWord
{{promise}}{{guard or none}}{{fired / quiet}}{{passed / fired}}{{caught N of N shapes / blind / jumpy / missing}}

{{the 3 claims checked, each still true or not}}

The old audit you were doing, next to the 4R audit

The old audit reads your files and tells you what the files say. The 4R audit also runs your workspace and plants mistakes. Every finding from today, and which audit could have found each 1.

Today's findingThe old auditThe 4R audit
{{finding}}{{finds this / sees the check exists / cannot see / not checked}}{{how the 4R audit found it}}

Since last time

{{better, worse or same per count, or "first audit, the baseline is set". Then every fix from last time: fixed, still open with its age, or worse}}

What this audit could not see

{{always at least 1 line}}