What problem does it solve? Debugging often devolves into guesswork, random code changes, and endless re-reading of the same files. This Skill enforces a disciplined root-cause investigation workflow that reproduces errors, gathers evidence, tests hypotheses systematically, and hands off a precise diagnosis — without ever touching the code itself. ## Core Features & Use Cases - Structured Hypothesis Testing: Forms exactly 2-3 ranked hypotheses, tests each with file:line evidence, and marks them CONFIRMED or RULED OUT, with a hard limit of 3 cycles before escalation. - Known Error Pattern Matching: Matches errors against a built-in catalog (MODULE_NOT_FOUND, ASYNC_DEADLOCK, PATH_MISMATCH, etc.) to skip hypothesis cycling for recognized patterns. - Scope Lock & Anti-Loop Guards: Locks investigation to the narrowest affected directory and detects evidence re-gathering loops to prevent analysis paralysis. - Use Case: A test fails with an unclear stack trace deep in execution. The Skill reproduces the error, backward-traces the invalid data to its origin, identifies the exact file:line root cause, and emits a Debug Report handed off to the fix skill. ## Quick Start Ask the agent to debug the failing test or error message you are seeing and produce a root cause report without changing any code.
