Use this skill whenever the user asks you to investigate, scope, scout, survey, or "look into" something before implementing. Trigger on phrases like "how hard would it be to...", "where would I even start with...", "I want to add X but I'm not sure what it touches", "do a spike", "do some recon", "don't write code yet", or any request where the scope is unclear. The job is to map the territory and surface unknowns — NOT to start implementing.
Use the `ketch` CLI for all web search, scraping, code search, and library doc lookups. Trigger whenever the user says "use ketch", asks to search the web, scrape a URL, look up OSS code examples, or fetch library docs — even if they don't mention ketch by name.
Use this skill whenever the user signals they want to merge a branch into the current one — phrases like "merge main", "merge into current branch", "bring in changes from X", "merge branch X", "sync with main". The job is to identify the source branch, show what would change (including a dry-run conflict check), and wait for approval before running any merge command.
Use this skill whenever the user signals they want to commit changes — phrases like "commit this", "let's commit", "make a commit", "prepare a commit", "ready to commit". The job is to show the user what would be staged and propose a commit message, then wait for approval. Do NOT run git add or git commit until the user explicitly approves.
Claude Code skills by vinhnguyen1211. These skills enforce a deliberate, preview-first workflow — Claude shows you what it's about to do and waits for explicit approval before taking any irreversible action.
Add the marketplace in Claude Code:
/plugin marketplace add vinhnguyen1211/skills
Then open /plugin, go to the Discover tab, and install the skills you want.
Goal: Understand the scope and surface unknowns before writing a single line of code.
Triggered when you say things like "how hard would it be to…", "where would I even start with…", "do a spike", or "don't write code yet." Claude reads the codebase, maps what exists, identifies risks, and writes a structured report — without touching any source files.
Expected output: A written report with six sections: goal restated, what exists today, a few approaches with tradeoffs, unknowns and assumptions, a recommendation, and the suggested next step. For small investigations this appears inline; for larger ones it's saved to exploration-<topic>.md.
Goal: Never commit without seeing exactly what's going in and why.
Triggered when you say things like "commit this", "let's commit", "make a commit", or "ready to commit." Claude runs git status and git diff to gather the current state, then proposes which files to stage and a real commit message — without running git add or git commit.
Expected output: A preview listing files to be staged (grouped by type), files intentionally left out, a draft commit message with subject and bullet-point body, and any flags (secrets, debug prints, unrelated changes mixed in). Claude waits for you to reply commit approved before running anything. Amend flows and multi-commit sessions are also handled — each step requires its own approval.
Goal: Never merge blindly — see exactly what's coming in before the history changes.
Triggered when you say things like "merge main", "merge into current branch", "bring in changes from X", or "sync with main." Claude identifies the source branch, then gathers facts with read-only git commands — without running git merge.
Expected output: A merge preview showing the current and source branches, incoming commits (listed or summarized if more than 10), files that would change, anything worth flagging (potential conflicts, wrong-direction merges, dirty working tree), and the proposed merge approach (standard merge commit, --ff-only, or --squash with reasoning). Claude waits for you to reply merge approved before running the merge.
Goal: Fast, stateless web search, scraping, code search, and library doc lookups via the ketch CLI.
Triggered when you say "use ketch", ask to search the web, scrape a URL, look up OSS code examples, or fetch library docs. Claude runs the appropriate ketch subcommand (search, scrape, crawl, code, or docs) and returns structured results.
Expected output: Results inline in chat — search snippets, scraped markdown, code matches, or library documentation — depending on the subcommand used.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub vinhnguyen1211/skills --plugin vinhnguyen1211-skillsPersistent file-based planning for AI coding agents. Crash-proof markdown plans (task_plan.md, findings.md, progress.md) that survive context loss and /clear, with an opt-in completion gate and multi-agent shared state. Manus-style. Works with Claude Code, Codex CLI, Cursor, Kiro, OpenCode and 60+ agents via the SKILL.md standard. Includes Arabic, German, Spanish, and Chinese (Simplified and Traditional).
Claude harness - A harness for solo developers (Vibecoders) to handle full-cycle contract development.
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Plugin-safe Claude Code distribution of Antigravity Awesome Skills with 1,561 supported skills.