dylanpulver
UserBattle-tested Claude Code skills I run every day — 20 skills in 5 installable plugins: session continuity, shipping, thinking tools, launch, and skills governance
Categories
Indexed Skills (22)
index-skills
Scan all installed plugins and skills across the system, generating a comprehensive skills-index.md. Use when asked to index skills, catalog what's installed, or refresh the skills registry.
repo-launch
Take a code repo public: gate the pitch, scrub secrets from git history, draft the Show HN post. Use for 'launch this repo', 'make this public', 'is this safe to open source'. NOT landing pages (site-launch) or novelty scans (prior-art).
prior-art
Scan the open-source / research landscape for projects doing what this repo does — is anyone ahead of us, has anyone had our ideas, and what are they doing that's worth porting in. Use when the user says 'prior art scan', 'who else is doing this', 'has anyone built this', 'are we still novel', 'competitive scan of this repo', or wants a novelty check on a project's distinctive ideas. Produces a pillar table, a novelty verdict, a steal list, and a watch list.
domain-scout
Find the best available, brandable domain for a concept: generate names, check whois across TLDs, screen trademark collisions, rank by brandability. Use for 'find a domain', 'is <domain> available', 'domain ideas for X', 'name this project'.
site-launch
Take a concept to a live landing page on a real domain: pick and buy the domain (Spaceship), build the page, deploy to Vercel, wire DNS + forms. Use for 'launch a site for X', 'set up a landing page + domain', 'buy a domain and deploy'.
recover-chats
Recover Claude Code chats after a crash across config dirs. Use for 'my vs code crashed', 'find all my open claude chats', 'recover my lost chats', or wanting labeled `claude --resume` commands across ~/.claude (and any additional config dirs).
codebase-tasks
Scan a codebase for TODOs, tech debt, and improvements — prioritize and optionally put up PRs for the quick wins
continuous-improvement
Audit a codebase for improvements, implement the best ones as PRs with auto-review and merge safety analysis. Loops in user for approval at key moments.
pr-fix
Review a GitHub PR, post findings as a comment, apply the fixes, push, and comment what was fixed and why
pr-safety
Assess the risk of merging a PR: what could go wrong, regressions, blast radius, rollback difficulty. Posts assessment as a PR comment.
pr-shepherd
Shepherd open PRs through a manual-merge workflow — assess risk, post consensus comments, propose merge order. Use when the user pastes a PR URL/number, says 'ok i merged X', 'whats next to merge', 'which of these are low risk'. Never merges.
blueprint
Design a new thing BEFORE building it — fuzzy idea to executable blueprint. Use for 'blueprint this', 'design this', 'architect this', 'scope this out', 'I have an idea for X'. NOT murky-problem diagnosis or plan stress-tests (premortem).
prompt-retro
Self-audit of Claude Code chat history — mine transcripts for repeated prompts, corrections, and friction, then propose skills/hooks/setup fixes. Use for 'prompt retro', 'run the chat-history audit', 'what patterns are in my prompting', or monthly.
prose-restraint
Writing discipline for substantive prose (docs, essays, READMEs): suppress over-explaining — state it once, show the implication, keep the omission. Use for 'prose-restraint', 'tighten this writing', 'stop over-explaining', 'trust the reader'.
second-opinion
Cross-check the current work against an independent model (Codex/GPT) as an ADVERSARIAL reviewer — it tries to break the work, not co-sign it. Use for 'second opinion', 'cross-check this', 'what would codex say', 'have another model review this', or before shipping something high-stakes/irreversible. NOT for routine diffs.
tool-triage
Triage a discovered tool/repo/skill/paper into PORT / ADOPT / WATCH / PASS for a named target, bloat cost counted; proposes first, executes on OK. Use for 'tool-triage X', 'should I adopt/port X', 'evaluate this tool'. ai-radar FINDS; this DECIDES.
skill-governance
Manage your Claude Code skills ecosystem. Use when asked to index skills, audit skill quality, list installed skills, check for skill conflicts, or manage the skills system.
handoff
Write or consume a standardized handoff document so work can continue in a fresh chat. Use for 'make a handoff', 'write a handoff doc', or when a session opens pointing at a handoff file to continue from it.
session-merge
Append this session's contribution as a structured block to a shared consolidation file, so several related chats can be merged into one picture. Use when the user says 'add what we made in this chat to <file>', 'combine this chat into the notes', 'session merge', or references combine-notes.
whereami
Answer 'where are we at and whats next' from a durable per-repo state ledger. Use for 'where are we at', 'whats next', 'tldr of where we stand', 'where did we leave off', 'take stock', resuming after a gap; update the ledger when work ships.
make-factory
Build a new factory — a repeatable pipeline turning an idea into one finished, quality-assured artifact type. Use for 'make a factory for X', 'this is a factory to make', '/make-factory', or when repeated multi-step output should become a factory.
research
Research factory — structured research to a decision-ready, sourced brief; searches the web AND your codebase/Drive/Slack/Gmail/prior briefs, flags where they disagree. Use for 'research X', 'look into X', 'should we use A or B', 'is there a tool for X'.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.