Audit, design, write, materially revise, or remove Tutti tests.
SkillScore95Maintained
Automatic
◆QA & Test Engineering pack
Last verified 9/8/2026
What it changes
Passed 20 of 21 test cases with the skill, and 11 of 21 without it.
Without skill50%
With skill95%
Checks it passed that the model alone missed
Specify reading docs/conventions/unit-testing.md
Include the header Test evidence in the evidence output
List Execution as one of the six authoring gate questions
and 22 more in the full report
10 fixed11 passed either way1 still failing0 broke
Pass-rate lift
+45 pp
Tokens
+54%
Agent turns
0%
Response time
-19%
Model gemini-3.6-flashMeasured 9/8/2026v121 of 22 cases passed with skill · 21 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 10 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 10
Prompt
Before starting a non-trivial test refactoring in the repository root, which specific convention documentation files must be consulted alongside AGENTS.md?
What was checked, without and with the skill
Checked
Without skill
With skill
Specify reading docs/conventions/unit-testing.md
Missed
Passed
Specify consulting docs/conventions/testing.md#validation-selection for validation scope
Missed
Passed
case-10 · gemini-3.6-flash · 9/8/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md160 linessingle file
Tutti Test Audit
Make every changed test earn its maintenance cost. A green test, a coverage
increase, or a larger test count is not a quality verdict.
Read the root and closest scoped AGENTS.md, then read
docs/conventions/unit-testing.md before a non-trivial test change or review.
Follow docs/conventions/testing.md#validation-selection for commands and
validation scope.
Establish the evidence map
Before editing, inspect:
the production owner, public entry point, relevant caller and callee;
sibling implementations that share the same invariant;
existing tests, fixtures, conformance suites, and repository checks;
the changed-aware and platform lane that will select the evidence;
relevant history for a regression or a suspicious existing test.
For Agent Host lifecycle semantics, also read
packages/agent/host/README.md and start with the conformance contract. Do not
reimplement lifecycle tests in an adapter.
Pass the authoring gate
Answer all six questions before adding or materially rewriting a test:
Protected contract: What observable behavior, invariant, compatibility
promise, or prior failure matters?
Credible failure: What plausible faulty implementation must make the
test fail?
Coverage gap: Why would existing evidence not catch that failure? Can an
existing table, scenario, or fixture express it without duplication?
Owner and observer: Which module owns the decision, and what is the
lowest boundary that can observe the risk without mocking it away?
Production seam: Does the test require an export, flag, wrapper, global,
or injection hook that no production caller needs? If so, move to the real
boundary or redesign the owner.
Execution: Which blocking lane selects the test, and which native OS must