Black-box optimization of design parameters using TuRBO or scipy. TRIGGER when the user wants to optimize, tune, size, sweep, or explore a design space to meet specs. This includes circuit sizing (W/L, bias, passives), finding optimal operating points, minimizing power-delay or noise-power tradeoffs, or any task where multiple parameters need to be searched to hit a target. Also trigger when the user says things like 'find the best sizing', 'help me tune this', 'run an optimization', 'what values give me the best FOM', or 'sweep these parameters to meet spec'. Do NOT trigger for single-variable parametric sweeps or analytical calculations.
Use when editing .lean files, debugging Lean 4 builds (type mismatch, sorry, failed to synthesize instance, axiom warnings, lake build errors), searching mathlib for lemmas, formalizing mathematics in Lean, or learning Lean 4 concepts. Also trigger when the user asks for help with Lean 4, mathlib, or lakefile. Do NOT trigger for Coq/Rocq, Agda, Isabelle, HOL4, Mizar, Idris, Megalodon, or other non-Lean theorem provers.
- 📁 hooks/
- 📁 references/
- 📁 scripts/
- 📄 config.json
- 📄 gotchas.md
- 📄 SKILL.md
Use this skill when a user asks to review a pull request for bugs, wants AI code review focused on correctness issues, or runs /bug-review. Trigger on PR review, bug finding, code review, "review this PR", "check for bugs", "find issues in this PR". This is a multi-pass review workflow with 5 parallel passes, majority voting, independent Opus validation, and resolution rate tracking. Also trigger on /bug-review:resolve to classify whether findings were fixed at merge time, and /bug-review:report for resolution rate stats. Even if the user just says "review this" while on a PR branch, trigger this skill.
Use when editing .lean files, debugging Lean 4 builds (type mismatch, sorry, failed to synthesize instance, axiom warnings, lake build errors), searching mathlib for lemmas, formalizing mathematics in Lean, or learning Lean 4 concepts. Also trigger when the user asks for help with Lean 4, mathlib, or lakefile. Do NOT trigger for Coq/Rocq, Agda, Isabelle, HOL4, Mizar, Idris, Megalodon, or other non-Lean theorem provers.
- 📄 code-analyzer.yaml
- 📄 pmd-ruleset.xml
- 📄 SKILL.md
Run Salesforce Code Analyzer on Apex, Trigger, Flow, or metadata files. Use after writing or modifying .cls, .trigger, .xml, or flow files. Also use when asked to scan, lint, or check code quality.
Use this skill at the START of a new Claude Code session to pick up where the previous session left off. Finds the most recent handoff document under `.claude/reports/handoff/`, reads it, then VERIFIES its claims against the live codebase (Read each referenced file, run `git status` / `git log --oneline`) before reporting findings and waiting for user instruction. Treats the handoff as a hypothesis, not fact — previous session may have been confused. Trigger on "catchup", "catch up", "이어서 작업", "이어서 해줘", "이전 세션 이어", "resume from handoff", "handoff 읽어", "handoff 참고해서 시작", "어제 작업 이어서", "지난번 이어서", "새 세션 준비", "prepare to continue", "pick up where we left off", "캐치업". Also trigger when the user starts a new session by referencing prior work ("어제 한 작업 이어서 하자", "계속 하자", "resume the auth refactor") and an `.claude/reports/handoff/` directory exists in the project. Do NOT trigger when the user wants to start a brand-new unrelated task, or when they want to read a specific named document (use Read directly).
Close the current work session with structured log, unsaved insights check, and open questions routing. Trigger when: (1) user says 'закрываем', 'закрывай сессию', 'всё на сегодня', 'финал', 'заканчиваем', 'close session', (2) user invokes /close-session, (3) user signals end of work session in any way. Do NOT trigger for: git commits, mid-session saves, 'сохрани' without session-closing intent, or 'закрой файл'.
Conduct deep, structured research on any topic — markets, products, features, competitors, problems, industries, regulations, or opportunities. Trigger whenever the user wants to research, explore, investigate, analyze, or understand something. This includes: entering new markets, evaluating products or services, exploring features, solving business problems, competitive landscapes, regulatory research, or any "I need to understand X" scenario. Also trigger for: "look into," "dig into," "explore whether," "map out the landscape," "what are the options for X," "how does Y work in practice." Trigger aggressively — if there's research intent, use this skill. Covers: market research, competitive analysis, feature exploration, regulatory deep-dives, product evaluation, problem-solving, opportunity assessment, landscape mapping. --- # Deep Research Skill ## Core Philosophy Most research is shallow. It covers what's easily Googlable, presents generic overviews, and stops at the surface. This skill exists to go far beyond that. **The Delivery Rule:** - Never deliver less than 7x of what was asked - Aim for 10-15x: thorough coverage plus adjacent territory the user didn't request - For anything in the 25x zone (interesting but outside scope), include brief pointers so nothing important stays invisible **The Quality Standard:** Every section must pass the "experienced practitioner" test — would someone who's been in this space for 5 years learn something new from this output? If not, go deeper. Surface-level overviews are never acceptable as a final deliverable. --- ## Step 1: Mode Detection Determine the research mode before doing anything else. ### Mode A: Guided (Interactive) **Trigger when:** The prompt is vague, broad, ambiguous, or could go in multiple very different directions. Or the user explicitly says "ask me questions first." **Process:** 1. Ask 3-5 sharp questions using the Question Decomposition technique (below) 2. Present a Research Plan and confirm 3. Execute
