最近新增
刚加入收藏的新技能，抢先试用。
differing-pair
jthorare
Portable Claude Code workflow toolkit: skills, config, repo bootstrapper, and MCP bridge for multi-agent communication
withInterview
jthorare
"Interview the user to clarify requirements before executing a prompt. Use when the user wants to refine a task through guided questions before implementation."
weeklyRetro
jthorare
Weekly retrospective — analyzes git history for per-person breakdowns, shipping streaks, test health trends, and generates actionable insights.
verify-web
jthorare
"Playwright-based self-verification of running web apps. Executes journeys through verification lenses (functional, visual, accessibility, error-state, responsive) and writes an evidence pack (pack.json + report.md) on every run."
review-change
taloncjones
Obtain one bounded independent development review of an implementation change.
herdr-orchestration
taloncjones
Use when Codex is asked to contribute UI/UX direction, bounded UI implementation, prose, or independent review to a Claude-led Herd task. This installed compatibility entrypoint points to the canonical shared workflow; it does not make Codex the Herd controller.
claude-spec-review
taloncjones
Obtain a bounded independent Claude review of one explicit frozen specification.
claude-plan-review
taloncjones
Obtain a bounded independent Claude review of one explicit frozen implementation plan.
writing-specs
taloncjones
Use after a design is approved in brainstorming, to write the PRD that the spec review gate checks and implementation works from.
writing-plans
taloncjones
Use only when the owner asks for a task-by-task implementation plan from a reviewed PRD; the default pipeline implements straight from the PRD.
wrap
taloncjones
Use when the user is ending a session and wants the exit state verified instead of being asked "anything else?". Trigger on "good to exit", "are we done", "wrap up", "closing out", "done for the day".
weekly
taloncjones
Compose my "Work Last Week / Work This Week" row for the weekly meeting Confluence page from my merged/open PRs and active Jira, and output it as a copyable block by default. Writes to my row only if I explicitly ask. Per-project config in ~/.claude/reconcile/projects.json.
todos
taloncjones
Use when capturing, listing, completing, or reviewing cross-session todos for a repo. Trigger when the user says "track this", "add a todo", "what todos are open", "mark that done", or wants a persistent backlog that survives across sessions (unlike the built-in TodoWrite tool, which is session-only).
ship
taloncjones
Use when a branch or PR is implementation-complete and the user wants it taken through the explicit merge gate and cleanup.
review-change
taloncjones
Obtain one bounded independent development review of an implementation change.
repo-recall
taloncjones
Use when answering questions about prior decisions, specs, plans, findings, todos, handoffs or session memory in the current repo -- "didn't we decide X", "what did the review say about Y", "find the spec/plan/todo about Z", "search our docs/notes/memory". Ranked SQLite FTS5 search over the repo's prose artifacts; not for code symbols (use LSP / grep).
reconcile
taloncjones
Reconcile my git/PR reality against Jira and fix tracking drift (untracked PRs, stale statuses, sprint hygiene) with propose-confirm-apply. Use when tidying my tracked work, before a sprint boundary, or when asked "what needs cleanup/action". Per-project config in ~/.claude/reconcile/projects.json.
security
musterd-io
The security role — the member who owns the team's application-security evidence (threat models, authorized scans, validated findings, private disclosure, the smallest defensible fix, and honest posture claims). Use when threat-modeling a system or a change, reviewing code or a workflow for security, running or reading a scanner, judging whether a finding is real and how severe, handling a leaked secret, reviewing a new dependency or the release pipeline, or writing anything public about security.
pr-status
taloncjones
'Use for a PR status table - every open PR in this repo with its head, CI, co-review verdict at head, bench evidence, body currency, draft state and what it waits on. Trigger on "pr status", "PR table", "status of the PRs", "what are the PRs waiting on", or /pr-status [pr ...] [--repo owner/name].'
planner
musterd-io
The planner role — the member who turns a pitched idea into work a team of agents and humans can build. Ask a few questions that change the plan, post one plan the team can accept in a minute, then declare the goal, open every lane at once and route each to the right member. Use when you hold the planner role and someone pitches an idea, or when asked to plan work for the team.
ui-ux
8-Sync-Dev
UI/UX và giao diện (design, redesign, frontend, landing, dashboard, component, a11y, responsive, animation, theme, UX copy) — điểm vào dễ gõ /ui-ux <việc>; chọn đúng skill chuyên sâu (impeccable là design system chuẩn của 8 Sync) rồi làm tới khi verify trên trình duyệt.
docs-that-catch-drift
musterd-io
Docs go stale for four specific reasons, and each has a structural fix — one doc one job, one fact one home, and "checker, not generator" — enforce the structure where it is mechanical and refuse to produce the prose where it is judgement. Ships a drift-checker for documented file trees with deliberately no --fix. Use when adopting docs conventions, when a doc has accumulated strikethroughs, or before writing a tool that generates documentation.
sysdesign
8-Sync-Dev
System design for a real change (new service, storage, queue, cache, scale, AI feature) using The System Design Primer as the reading source — pick the sections, run the 4-step method with back-of-envelope numbers, write the ADR. Use when a task adds a component, changes data flow or must survive more load.
dated-and-falsifiable
musterd-io
A belief carries the date it was observed and the observation that would overturn it — plus the twelve named ways a green check or a quiet instrument lies to you, each with a repair. Use when writing down what you learned, when a check passes and you are about to believe it, when a tool reports nothing, or when a team keeps re-discovering the same thing.