Startup foundation · 01
Assumption Ledger
A planned ledger that makes a venture's testable assumptions visible and ranks them by stakes if wrong and cost to test.
Collection 01 · Startup foundations
Nine Planned capability records for turning assumptions, experiments, economics, delegation, and governance into work a founder can inspect.
Nine recurring problems
Each record names a business need, declares a proposed operating contract, and makes the remaining implementation and review work visible.
Startup foundation · 01
A planned ledger that makes a venture's testable assumptions visible and ranks them by stakes if wrong and cost to test.
Startup foundation · 02
A planned one-page diagnostic for identifying a venture's most binding liability of newness and one near-term response.
Startup foundation · 03
A planned adversarial pre-mortem that compares founder-identified risks with failure modes surfaced through four hostile perspectives.
Startup foundation · 04
A planned workflow for creating the smallest test-ready artifact that can probe one behavioral claim.
Startup foundation · 05
A planned unit-economics model that tags every line item by what is known, variable, or still unknown.
Startup foundation · 06
A planned simulation workflow for showing a distribution of contingency runway rather than a single confident forecast.
Startup foundation · 07
A planned map for assigning venture work across founders, hires, contractors, partners, and agents using task characteristics.
Startup foundation · 08
A planned operational map linking task assignments to oversight, recovery, and escalation paths.
Startup foundation · 09
A planned kit for drafting a disclosure architecture, data provenance map, and audit-agent specification that can constrain venture choices.
Build discipline
Assumption Ledger is the proposed harness proof. Only after its frozen bytes pass a different-vendor gate does Liability of Newness become the first composition proof. Both records remain Planned today.
Define the founder question and the cost of a wrong answer before specifying agent behavior.
State the inputs, outputs, permissions, failure states, and human gates the capability requires.
Keep every record Planned until the runnable artifact and review evidence can be inspected together.
One capability map