Problem
No serious counterparty will escrow funds in unaudited contracts, and most
grant programmes ask about audit status. Audits are also expensive and slow, and
arriving unprepared wastes a large fraction of the engagement on things the team
could have supplied.
What to do
- Write
docs/AUDIT_SCOPE.md: which contracts, which commit, what is in and
out of scope, and the trust assumptions the auditor should take as given.
- Assemble the readiness package: threat model, invariants, architecture notes,
known issues, and full test coverage figures.
- Freeze the interface before the engagement — auditing a moving target wastes
the budget.
- Budget for a fix-and-review round; the first report is not the end.
- Plan to publish the report, including unresolved findings and why they were
accepted.
Acceptance criteria
Notes
Publishing the report including unfixed findings builds more trust than a clean
summary. Reviewers who have read a few audits know there is no such thing as
zero findings.
Problem
No serious counterparty will escrow funds in unaudited contracts, and most
grant programmes ask about audit status. Audits are also expensive and slow, and
arriving unprepared wastes a large fraction of the engagement on things the team
could have supplied.
What to do
docs/AUDIT_SCOPE.md: which contracts, which commit, what is in andout of scope, and the trust assumptions the auditor should take as given.
known issues, and full test coverage figures.
the budget.
accepted.
Acceptance criteria
Notes
Publishing the report including unfixed findings builds more trust than a clean
summary. Reviewers who have read a few audits know there is no such thing as
zero findings.