qingyi

About this skill

What's included (14 skills in this plugin)

• commit-push

  This skill should be used when the user wants to commit AND push changes automatically without any confirmation prompts. Fully autonomous: stages, commits, and pushes. ACTION TRIGGERS: - "commit and push", "push changes", "save and push", "quick commit" - "push it", "ship it", "send it", "push everything" - "auto commit", "just push", "commit push" NATURAL LANGUAGE PATTERNS: - "commit and push" + code/changes context - "push" + all/everything/changes - "ship it" / "send it up" / "push it up" - "quick commit and push"

• commit

  This skill should be used when the user wants to commit code changes to git and optionally push to a remote. ACTION TRIGGERS: - "commit", "git commit", "commit this", "commit my work" - "save changes", "create a commit", "conventional commit" NATURAL LANGUAGE PATTERNS: - "commit" + code/changes context - "save my changes" / "commit what I have" - "done, commit" / "looks good, commit" NOTE: For "commit and push" or "push changes", use the commit-push skill instead.

• debug-code

  Use when the user encounters bugs, errors, crashes, or unexpected behavior. ACTION TRIGGERS: - "debug", "fix bug", "fix error", "troubleshoot", "investigate error" - "not working", "broken", "crashed", "why is this failing" - "trace the error", "find the bug", "root cause" NATURAL LANGUAGE PATTERNS: - "I'm getting" + error/exception/traceback - "help me debug" / "help me fix" - "here's the error" / "here's the log" - "the log says" / "the output shows" PROACTIVE TRIGGERS (fire even without explicit debug request): - User pastes error messages, stack traces, tracebacks, or error logs - User pastes log lines with ERROR, FATAL, Exception, Traceback, panic - User reports HTTP errors, DB errors, test failures, or compile errors - User describes symptoms: "returns null", "blank page", "hangs", "freezes"

• excel-edit

  This skill should be used when the user wants to WRITE, FORMAT, STYLE, or DESIGN Excel files. FILE-BASED TRIGGERS (highest priority): - Any mention of ".xlsx" or ".xls" files WITH edit/create intent - "the excel file", "this excel" + modification context ACTION TRIGGERS: - EDIT: "edit excel", "modify spreadsheet", "update cells", "change values" - CREATE: "create excel", "new spreadsheet", "make xlsx", "build a table" - FORMAT: "format excel", "style cells", "add colors", "professional formatting", "beautify spreadsheet", "eye-catching", "modern look" - STRUCTURE: "add worksheet", "merge cells", "add chart", "pivot table", "add rows", "delete columns", "insert data", "freeze panes" - ORGANIZE: "reorganize excel", "restructure spreadsheet", "clean up excel", "sort spreadsheet", "rearrange columns" NATURAL LANGUAGE PATTERNS: - "make the excel" + improvement verb - "add" / "change" / "update" + excel/spreadsheet context - File path ending in .xlsx or .xls + modification request

• excel-read

  This skill should be used when the user wants to VIEW, ANALYZE, or EXTRACT data from Excel files. FILE-BASED TRIGGERS (highest priority): - Any mention of ".xlsx" or ".xls" files WITH read intent - "the excel file", "this excel", "my spreadsheet" + viewing context ACTION TRIGGERS: - "read excel", "open xlsx", "view spreadsheet" - "show excel data", "what's in the excel", "display spreadsheet" - "check the excel", "look at the spreadsheet" - "analyze spreadsheet", "extract data from excel" - "count rows", "find values", "list worksheets" - "summarize the excel", "what sheets are there" - "parse excel", "import from excel", "convert excel to" NATURAL LANGUAGE PATTERNS: - "show me" + excel/spreadsheet - "what does the excel contain" - "read" + file path ending in .xlsx or .xls

• fetch-notion-tasks

  Fetches tasks from a Notion database using multi-query search to work around Notion MCP's semantic search limitations, optionally rewrites empty descriptions, and executes all verified tasks concurrently — each in its own git worktree under `.claude/worktrees/` — before handing off to `merge-worktree` to land the changes. This skill should be used when the user asks to "fetch tasks", "get notion tasks", "check tasks", "show tasks", "list tasks", "what tasks are in progress", "pull tasks from notion", "check notion board", "rewrite task descriptions", "plan tasks", "plan implementation", "implement notion tasks", "what needs to be done", or mentions task status from Notion.

• improve-instruction

  This skill should be used when the user asks to "improve instructions", "trim project instructions", "reduce CLAUDE.md tokens", "audit instruction files", "clean up project instructions", "review and fix CLAUDE.md", "remove outdated sections from instructions", "optimize context files", "reduce context size", "fix inaccuracies in instructions", or "improve CLAUDE.md". Provides a structured workflow for auditing, correcting, and pruning Claude Code instruction files to reduce per-conversation token cost without losing actionable content.

• merge-worktree-no-commit

  This skill should be used when the user wants to land worktree changes into the main branch WITHOUT committing them in the worktree first. Uses git stash through the shared .git directory to transport uncommitted edits; existing commits on the worktree branch are still merged normally. The source worktree and its branch are left in place after a successful transfer. ACTION TRIGGERS: - "merge worktree without commit", "merge worktree without committing" - "merge worktree no commit", "merge worktree skip commit" - "transport worktree changes", "move worktree changes to main" - "merge worktree wip", "land uncommitted worktree changes" - "merge worktree but keep changes uncommitted" NATURAL LANGUAGE PATTERNS: - "merge" + "worktree(s)" + ("without committing" / "no commit" / "skip commit") - "bring" / "move" / "transfer" + worktree changes + main/master - "I don't want to commit in the worktree" + merge context - "pull" / "pick up" + worktree edits + main + uncommitted

• merge-worktree

  This skill should be used when the user wants to commit, merge, and clean up git worktrees back into the main branch. Handles uncommitted worktree changes, merges all branches, removes worktrees, deletes feature branches, optionally pushes, and summarizes all changes. ACTION TRIGGERS: - "merge worktrees", "merge worktree", "merge all worktrees" - "commit worktrees and merge", "finish worktrees", "land worktrees" - "clean up worktrees", "merge branches back", "consolidate worktrees" - "merge worktree changes", "close worktrees", "wrap up worktrees" NATURAL LANGUAGE PATTERNS: - "merge" + "worktree(s)" / "branch(es)" + optional "push" - "finish" / "land" / "close" / "wrap up" + "worktree(s)" - "commit all worktrees and merge into master/main" - "clean up worktrees and push"

• obsidian-init

  This skill should be used when the user wants to create an Obsidian vault that documents a codebase as a graph-linked, atomic-note knowledge base. Scope: greenfield initialization only — NOT for updating, refreshing, or re-syncing an already-populated vault. ACTION TRIGGERS: - "create obsidian vault", "init obsidian", "scaffold obsidian vault" - "document codebase in obsidian", "make obsidian knowledge base" - "obsidian vault for project", "set up obsidian", "build obsidian vault" - "obsidian knowledge vault", "vault for documentation", "obsidian-style docs" NATURAL LANGUAGE PATTERNS: - "obsidian vault" + codebase/project/docs context - "set up" / "build" / "create" / "init" + obsidian + vault - "document" + codebase + obsidian - "map my codebase" / "second brain for this repo" / "zettelkasten for this codebase" - "build a wiki for" + repo / project / codebase - "link notes about this project" PROACTIVE TRIGGERS (fire even without naming Obsidian): - User wants a navigable knowledge graph / wiki / MOC / Map of Content for a repo - User mentions Logseq, Zettelkasten, or atomic-note documentation for a codebase - User has created an empty `obsidian/` folder and wants it pre-populated DO NOT fire when: - User wants to update / refresh / re-sync an existing populated vault - User mentions only "knowledge base" or "knowledge vault" without Obsidian context (could be RAG / embeddings work)

• remove-deadcode

  This skill should be used when the user asks to "scan for dead code", "remove dead code", "find unused code", "clean up unused exports", "find unused dependencies", "remove unused imports", "prune dead files", "find orphaned files", "check for unused functions", or "clean up barrel exports". Also triggers on mentions of "dead code", "unused code", "orphaned code", or "stale code". Provides a language-agnostic phased workflow: parallel exploration with grep-verified evidence, false positive awareness for framework patterns, cascade detection, safe removal ordering, and post-removal verification. Works with any language or stack.

• rewrite-history

  This skill should be used when the user wants to edit, rewrite, or fix old git commit messages using git-filter-repo. ACTION TRIGGERS: - "rewrite history", "rewrite commit", "reword commit", "edit commit message" - "fix commit message", "change commit message", "rename commit" - "filter-repo", "bulk edit commits", "replace in commit messages" - "fix old commit", "update commit messages", "clean up history" NATURAL LANGUAGE PATTERNS: - "rewrite" / "reword" / "edit" / "fix" + commit message context - "change the message on" + commit hash/description - "find and replace" + commit messages - "bulk update" + commit messages / history - "typo in commit" / "wrong commit message" NOTE: For amending only the most recent commit message, use `git commit --amend` directly. This skill targets older commits that require history rewriting.

• sentry-deploy

  This skill should be used when the user asks to "deploy Sentry", "set up Sentry", "add Sentry to my project", "integrate Sentry error tracking", "create a Sentry project", "configure Sentry monitoring", "add Sentry error reporting", "set up Sentry profiling", "configure Sentry logging", or "set up Sentry DSN". Covers end-to-end Sentry deployment: project creation via MCP tools, SDK installation, and configuration of error tracing, performance tracing, profiling, and logging for any language or framework.

• setup-permission

  This skill should be used when the user wants to configure Claude Code permission rules (allow/deny entries in settings.json). ACTION TRIGGERS: - "setup permissions", "configure permissions", "allow tools", "permission rules" - "settings.json", "setup allow rules", "generate permissions", "auto-allow" - "scan for commands", "discover tools", "permission config" - "stop asking for permission", "auto-approve", "tired of approving" NATURAL LANGUAGE PATTERNS: - "set up permissions" + project context - "allow" / "whitelist" + tools/commands - "generate settings.json rules" - "what commands should I allow" - "scan my project for permission rules" - "stop prompting me" / "auto-approve everything" - "trust these tools" / "reduce permission prompts" NOTE: Deny rules are ONLY generated when the user explicitly asks for them (e.g., "add deny rules", "block commands", "restrict tools"). Otherwise, only allow rules are produced.