Make the case. Show the evidence.

What evidence would justify this particular release claim?

Make your own investigation

Build a claim graph with your own evidence, failed tests and dependencies. Defend a bounded release or a decision to stop.

Build dependency links by claim ID. Surface unknown links, cycles, failed tests, missing evidence and unresolved defeaters.

Case cards

Cut these cards apart. Use the original example first, then let a partner supply a changed case. Keep sealed/final cards with the teacher until the learner commits a rule.

Case 1

Claim ID: C1

Claim: Answers stay inside supplied evidence

Evidence: 20 answerability cases

Passes: 18

Tests: 20

Depends on:

Defeater: Two unsupported answers

My prediction and reason:

Case 2

Claim ID: C2

Claim: External send is unavailable

Evidence: Permission denial tests

Passes: 6

Tests: 6

Depends on:

Defeater:

My prediction and reason:

Case 3

Claim ID: C3

Claim: Pilot can proceed within scope

Evidence: Review record

Passes: 1

Tests: 1

Depends on: C1,C2

Defeater:

My prediction and reason:

What this cannot establish

An internally consistent graph does not establish evidence relevance or truth; independent review remains necessary.

Open the release brief and test records.