Trace code execution step-by-step for any input scenario

A Logic-Lens skill for execution explain: producing a step-by-step, interprocedural execution trace of what a specific piece of code actually does for a given input, including name resolution and type transitions. It uses lazy loading of shared reference docs (common.md for language, report header variants, scope routing, and loading budget; logic-explain-guide.md loaded step by step; semiformal-guide.md, semiformal-checklist.md, and report-template.md loaded on demand). Notably, logic-risks.md is intentionally skipped - this skill does not produce L-code findings and Remedy is explicitly out of scope. If the trace reveals a bug, the skill stops and recommends logic-review or logic-locate instead, and when handing off it preserves the premises already established and trace steps already completed under a "Partial trace context (carry into next skill):" heading so the user can pass that work directly into the follow-on skill rather than losing it. The process runs six steps: Step 0 detects language and confirms a single function plus a single input scenario is in scope, handing off to logic-review if the user wants bug-finding without a concrete scenario. Step 1 names the entry point function, the input scenario, and what the user is trying to understand. Step 2 builds premises by resolving every non-obvious name, stating the types of key variables at entry, and noting any global or module state accessed. Step 3 produces a numbered, interprocedural, active-voice trace, crossing function boundaries whenever relevant to the user's scenario while staying scenario-bound - it does not branch into alternative paths unless they explain the user's confusion. Step 4 highlights non-obvious behavior: name resolutions, implicit coercions, and hidden side effects - the gotchas a casual reader would miss. Step 5 summarizes actual behavior versus assumed behavior in one sentence each, which is described as the core value delivered to the user. The mode line in the report reads "Execution Explain" (执行解释 in Chinese), and because the mode is descriptive rather than evaluative, the report omits the Logic Score, Fault Confidence, and Verdict lines entirely.

Use it when a user is confused about what code does for a specific input and asks to be walked through it - "walk me through this", or similar - for a single function and a single scenario. Hand off to logic-review instead if the user wants bug-finding without naming a concrete input scenario; hand off to logic-review or logic-locate if the trace itself surfaces a bug, since Remedy is out of scope here.

Input is one function plus one concrete input scenario. Output is a numbered, interprocedural execution trace with resolved premises, a highlighted-gotchas section, and a one-sentence-each actual-versus-assumed summary - with no Logic Score, Fault Confidence, or Verdict line, and no L-code findings.

Part of the Logic-Lens skill family, sharing ../_shared/common.md, ../_shared/semiformal-guide.md, ../_shared/semiformal-checklist.md, and ../_shared/report-template.md with sibling skills logic-review, logic-locate, and logic-diff, deliberately not loading logic-risks.md, and handing off partial trace context to logic-review or logic-locate when a bug surfaces mid-trace.

Developers who are confused about why code behaves a certain way for a specific input and need a descriptive, step-by-step walkthrough rather than a bug hunt or equivalence check.