Not all evidence is equal, and pretending otherwise is how portfolios lie. Every claim here is graded T1, T2, or T3 — and the grade is displayed, including when it is the weakest one.
T1LIVE REPO / CI REPORT / DEPLOYED DEMO
Evidence that runs somewhere you can reach without me: a live repository, a CI report, a deployed demo. T1 survives my laptop being off.
WHAT QUALIFIES
- A public repo you can clone and run
- A CI workflow with green (or honestly red) runs attached to real commits
- A production deploy you can click through yourself
ON THIS SITE
Case studies badged LIVE, and ledger rows whose artifact is a repo, workflow file, or public deploy.
T2LOCAL PROOF / TERMINAL OUTPUT / DIAGRAM / RUNBOOK
Evidence that exists as an artifact on disk — terminal output, a generated report, flow files, a runbook — but was produced locally. Real, inspectable, not yet independently re-runnable.
WHAT QUALIFIES
- A machine-generated report (never hand-edited) checked into a repo
- Terminal gate output or a coverage summary from a recorded run
- Generated files on disk — test flows, ledgers, diagrams — that back a claim
ON THIS SITE
Case studies badged LOCAL PROOF, and ledger rows whose artifact is a report, script output, or on-disk directory.
T3WRITTEN EXPLANATION / PLANNED WORK
A written explanation or planned work. No artifact yet — the claim exists only as prose. T3 is allowed on this site under one condition: it is labeled T3.
WHAT QUALIFIES
- A recorded result whose artifact must be regenerated before it can be quoted
- A design, plan, or runbook for work that has not run yet
- Anything where the honest answer to "show me" is a document, not an output
ON THIS SITE
Ledger rows marked T3 — like a device-cert number that honestly reads 0/256 until re-run — and every "improve next" note on the case studies.