skeptic

Root cause analysis validation involves generating and validating competing alternative theories with supporting evidence. This skill challenges primary hypotheses by searching the codebase to find corroborating evidence for alternative incident diagnoses.

Hallucination detection against code citations verifies claims by comparing them against actual code. This skill flags unsupported statements and checks whether cited code evidence genuinely supports the primary hypothesis or proposed alternative theories.

Yes, evidence validation can find alternative theories for a suspected database issue. The skill searches the codebase for supporting evidence, potentially proposing a faulty caching layer or other alternative root causes backed by relevant code files.

Challenging hypotheses with alternative theories improves incident diagnosis by reducing bias and ensuring robustness. By generating competing root cause theories and validating them with codebase evidence, you avoid prematurely accepting an incorrect primary diagnosis.

Yes, code search is required for evidence gathering when validating alternative root cause theories. The skill depends on searching the codebase to find supporting evidence for each alternative and verifying that cited files actually substantiate the claims.

The limitation of generating alternative theories for incident diagnosis is that the quality depends entirely on available codebase evidence. If the code search lacks sufficient context or relevant files, the skill cannot produce well-supported competing root cause hypotheses.