Private learning · Trusted knowledge · Emergent value
Let your AI work,
work for you.
Your AI work contains talent signals no resume shows — and most of them you can't see from inside the work. Capture meaningful work from Claude, Codex, and Cursor into one logbook, learn from the patterns it reveals, and share proof when it counts.
plugins capture only when you asklogbook private by defaultsharing yours to revoke
↓ Three steps, no extra workflowLearning to ask the second question
Debugging a flaky OAuth callback — the second prompt asked what evidence would distinguish the three causes.
Work type what you were doing
Skills how you worked
Domain where it applies
Tools what you used
Work type first — skills measured in context, not guessed.
01 · How it works
Three steps. No extra workflow.
01 / Connect a plugin
Claude Code, Codex, or Cursor. Install it, connect your account, and keep working the way you already do.
02 / Capture what mattered
When a session is worth keeping, say so. It goes into your logbook with the project it belongs to. Nothing gets captured unless you ask.
03 / Review and share
Watch your record build. When it matters — a screening, a client — share proof from a link you control and can turn off.
02 · Your tools
Begin where you work.
Claude Code Available
Install from the public marketplace. Two commands, then capture on request.
Setup steps →claude.ai & ChatGPT In review
Directory listings under review. The custom-connector path is documented today.
Connector docs →The Chrome browser extension records your AI conversations automatically, building your record in the background. You choose what gets analyzed, and you can delete anything.
03 · The practice
Get better at directing AI.
Spot your patterns
Where the second question paid off. Where you accepted too early. The record shows how you actually work with AI — not how you think you do.
Sharpen your context
Most sessions go sideways because the AI never saw the right context. Your logbook lets you reread a session from outside it — that's where the thin spots show, so the next one starts smarter.
One record, every platform
Claude, Codex, Cursor — your record follows you. What you learn in one tool doesn't get left behind in it.
04 · What it shows
The skills you can't put on a resume.
Patterns you can't see yourself
How you reason, check your work, and recover from dead ends shows up across many sessions. You can't see it from inside one. Your logbook can.
A record you learn from
Go back through your own history. Find the session where something clicked, see what you did differently, and leave yourself a note.
Proof when it counts
Nothing is shared until you share it. When you do, anyone can check your certification at a public link — and it can't be quietly edited afterward.
05 · Your controls
Private by default.
Your conversations are encrypted and stored securely. You control what gets captured, what gets analyzed, and what gets shared. Recruiters see nothing unless you share a link — and you can delete any entry, any time. If you ever want your work to help research, that's a separate switch, off until you turn it on.
06 · The other side
For recruiters
You see what a candidate chooses to share: real work, checkable in one click. No scraped profiles, no inferred scores. Access is opening in invite-only waves.
vs.proof.v1 · whole profile Illustrative
Proof of Work
ValueSignal member · 23 scored sessions
app.valuesignal.ai/verify/…payload hash ·
sha256 9c41 f2aa …valid through the certificate's stated date