分析與評估 (Analysis & Estimation)

Common Lab (experimental): 15 TypeSafe Jev lab skills — a Jev request router (suggest-jev), Jev setup (init-jev), yes/no gate (jev-gate), list pick (jev-pick, incl. a before-stopping check), rubric score (jev-score), Jev opportunity finder (jever), jev-browser (the dev-browser CLI with Jev element picking), an ask-the-user check (is-truely-need-to-ask-user-jev), and Jev overlays that add Jev readings to define-goal, strategic-advance, review, think, wait-what, wayfinder, and cold-estimation (*-jever). Graduated skills now live in the stable plugins: grilling, research, prototype, wait-what, wayfinder in common; think, review in baransu. The Lab estimate, define-goal and better-prompts variants were retired in 0.9.0, the Lab strategic-advance variant in 0.10.0; 0.19.0 retired ui-jever, seal-jever, show-me-jever, none-stop-jever, and both Lab delegate variants with the executor agent; use analysis-estimation:cold-estimation, common:define-goal, common:better-prompts, common:strategic-advance, common:delegate, baransu:ui / baransu:draw, baransu:review and baransu:show-me; 0.20.0 retired contract-jever with the upstream contract skill — implement directly under red/green TDD and re-verify with baransu:review. Invoke as /common-lab:<skill>. Enabled by default once installed; disable with `claude plugin disable common-lab` if unwanted.

15 skills

Everyday dev toolkit: prompt tuning, goal setting, delegation, long-running strategic execution, wayfinder decision maps, and plain-language or visual explainers.

9 skills · 2 agents

Claude Code mods (Claude only; enabled by default; no slash commands): recall — typing # in the prompt box or the 回想 button opens a band above the prompt to search past Claude Code, desktop and Cowork conversations; pressing one puts its #chat:<id> token in the prompt box, and sending the prompt attaches that conversation as quoted context; bar — a two-rail status line above the prompt: folder, model and animated ctx, 5h and 7d gauges, then buttons for 回想, diff, artifacts, 看不懂 (/common:wait-what), 畫給我看 (/baransu:show-me) and 下一步 (next-step suggestions put in the prompt box as a draft); the terminal and the desktop each draw it their own way: the terminal with neon gauges and chips, the desktop with native buttons and SVG gauges and without diff and artifacts; 看不懂 and 畫給我看 fall back to a project's own wait-what and show-me skills, and a project adds its own buttons in .claude/common-mod.json.

Talk like caveman. Cut filler. Keep technical details explicit.

22 skills · 1 command · 3 agents · hooks

Persistent memory for Claude Code: the next session starts where the last one left off.

29 skills · 1 MCP server · hooks

tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.

1 skill

AI-powered codebase understanding — analyze, visualize, and explain any project

9 skills · 10 agents · hooks

Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications

5 skills · hooks

Cross-session memory and token savings for coding agents.

6 skills · 1 agent · 1 MCP server · hooks