duo-divergence-skill

Duo-mode prompt validation detects intentional mismatches between SKILL.md and codex/prompt.md instruction bodies, verifying that your AI workflow harness correctly flags divergence when expected and does not falsely pass consistency checks.

You test instruction consistency by intentionally providing mismatched SKILL.md and codex/prompt.md content as a test fixture, then running the skill to confirm that duo-mode divergence detection triggers when expected and does not falsely pass.

Your duo-mode system likely fails to catch mismatched prompt behaviors because it lacks intentional divergence testing fixtures that deliberately provide different instruction bodies to trigger and verify divergence detection in your regression checks.

No specific dependencies or components are required to run prompt divergence regression checks. You only need to define a name and description in the SKILL.md frontmatter and deliberately include a mismatch in the instruction bodies.

Use a divergence testing harness for prompt validation when evaluating a duo-mode implementation, running regression checks to ensure divergence detection triggers when expected and does not falsely pass consistency validation.