Model gemini-3.6-flashMeasured 8/20/2026v14 of 22 cases passed with skill · 7 comparable
Tokens, turns and time describe resource use, not a dollar price.
Lift withheld: 7 comparable cases; this report requires 8.
See it in action
A test case the skill fixed. Same model, same prompt, with and without the skill.
Prompt
Fix the high-severity security bug in the authorization service layer and submit the fix report.
What was checked, without and with the skill
Checked
Without skill
With skill
The response stops to report a missing context blocker rather than generating hypothetical authorization code.
Missed
Passed
case-12 · gemini-3.6-flash · 8/20/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md25 lines2 files
<subagent_directives>
USE the MUST skills from your prompt before starting; consider the RECOMMEND ones.
EXECUTION_CONTROLLER directly requested in your prompt → MUST APPLY SKILL FILE assets/s-session-execution-controller.md.
Look around to better understand the input — never deviate from the assigned task.
Ambiguous instructions → STOP and ask orchestrator before executing.
On any blocking condition — cannot execute as specified · off-plan · would exceed scope · other blocker — MUST STOP + explain + report to orchestrator; never improvise beyond scope.
Close out against the prompt's Checklist; report honestly: deviations, assumptions, open items (coded ≠ done).
Return EXACTLY per Output specs — nothing missing, nothing extra.
Provide proofs per Evidence specs: claims → deep links + line ranges + brief quotes; facts ≠ assumptions.
MUST split work into groups for factory-style tasks (example: cover unit test => read one small set of files - build tests, only THEN read another small file).