Multi-agent adversarial-verification orchestration for high-confidence conclusions. Fan-out finders, then verify every finding through a three-prism panel (exploitability / correctness / refutation) that defaults to disbelief, gate fixes behind load-bearing proof tests that catch agents who falsely claim "done/fixed", and roll out behind a build-first exit-code guard.
SkillScore86Maintained
Automatic
◆Security & AppSec pack
Last verified 8/24/2026
What it changes
Passed 18 of 21 test cases with the skill, and 11 of 21 without it.
Without skill55%
With skill86%
Checks it passed that the model alone missed
Defers verification until all discovery phase findings are collected.
Requires a concurrent execution test that passes with the lock present.
Requires that the test fails if the authorization check code is reverted.
and 6 more in the full report
8 fixed11 passed either way2 still failing1 broke
Pass-rate lift
+32 pp
Tokens
+30%
Agent turns
0%
Response time
-23%
Model gemini-3.6-flashMeasured 8/24/2026v119 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 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 7
Prompt
We are orchestrating 5 parallel implementer agents to apply fixes across 10 shared source code files in a single git repository directory at the same time. What workspace isolation rule must be enforced?
What was checked, without and with the skill
Checked
Without skill
With skill
Mandates git worktree isolation for concurrent writing agents.
Passed
Passed
Restricts workspace modifications to a single implementer when worktree isolation is absent.
Missed
Passed
case-10 · gemini-3.6-flash · 8/24/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.