Applies rigorous adversarial analysis to generate, critique, fix, and consolidate solutions for any problem (technical or non-technical).
SkillScore23Maintained
Automatic
◆AI Engineer pack
Last verified 8/18/2026
What it changes
Passed 5 of 21 test cases with the skill, and 3 of 21 without it.
Without skill14%
With skill23%
Checks it passed that the model alone missed
Executes scripts/export_analysis.py upon finishing.
Exports the output using scripts/export_analysis.py.
Calls scripts/export_analysis.py to export the markdown report.
and 3 more in the full report
4 fixed1 passed either way15 still failing2 broke
Pass-rate lift
+9 pp
Tokens
+15%
Agent turns
0%
Response time
+7%
Model gemini-3.6-flashMeasured 8/18/2026v15 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 4 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 4
Prompt
We want to secure our GitHub Actions CI/CD pipelines against supply chain attack vectors and malicious PR runner injection. We might just require manual approvals for all workflow runs. Walk through a multi-phase critique of candidate pipeline security frameworks, develop verified fixes for weaknesses, rank options by feasibility, risk, impact, and resources, provide a final path, and export the report.
What was checked, without and with the skill
Checked
Without skill
With skill
Ranks final options by Feasibility, Impact, Risk Level, and Resource Requirements.
Passed
Passed
Exports the output using scripts/export_analysis.py.
Missed
Passed
case-10 · gemini-3.6-flash · 8/18/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md97 lines2 filesincludes scripts
Adversarial Prompting
This skill applies a structured adversarial methodology to problem-solving by generating multiple solutions, rigorously critiquing each for weaknesses, developing fixes, validating those fixes, and consolidating into ranked recommendations. The approach forces deep analysis of failure modes, edge cases, and unintended consequences before committing to a solution.