Claude Code: Development & Testing Skills

Evidence-driven investigation for network, streaming, and protocol-layer bugs. Use when debugging connection resets (ECONNRESET, HTTP/2 RSTSTREAM, INTERNALERROR), SSE or long-polling stalls, fixed-time connection drops, CDN/proxy/CGNAT idle timeouts, or any incident where symptoms do not match the obvious cause. Applies falsification-first methodology — layered isolation experiments to pin down the responsible network layer, env-gated runtime instrumentation for non-invasive observation, and counter-review agent teams to challenge single-cause assumptions. Strongly trigger on "socket closed unexpectedly", "stream interrupted", "ECONNRESET", "HTTP/2 INTERNALERROR", "fails after N seconds", "works sometimes but not always", "upstream silent for X seconds", or any scenario where the investigator might jump to conclusions before evidence. Generalizes to any multi-layer system investigation where assumption-first thinking is the failure mode.

Use tunnel-doctor in Claude Code when Tailscale partially works but SSH, browser, git, or Docker paths fail because proxy, route, or double-tunneling conflicts need a structured diagnosis.

Export and sync Douban (豆瓣) book/movie/music/game collections to local CSV files via Frodo API.

Converts any Claude Code skills repository into an official plugin marketplace. Analyzes existing skills, generates .claude-plugin/marketplace.json conforming to the Anthropic spec, validates with claude plugin validate, tests real installation, and creates a PR to the upstream repo. Encodes hard-won anti-patterns from real marketplace development (schema traps, version semantics, description pitfalls). Use when the user mentions: marketplace, plugin support, one-click install, marketplace.json, plugin distribution, auto-update, or wants a skills repo installable via claude plugin install. Also trigger when the user has a skills repo and asks about packaging, distribution, or making it installable.

Reviews and improves Claude Code skills against official best practices. Supports three modes - self-review (validate your own skills), external review (evaluate others' skills), and auto-PR (fork, improve, submit). Use when checking skill quality, reviewing skill repositories, or contributing improvements to open-source skills.

Narrative-first slide deck creation. Guides users through structured narrative design (ABCDEFG model), then delegates visual generation to baoyu-slide-deck. Triggers on "create slides", "make a presentation", "generate deck", "slide deck", "PPT", or when user needs to turn content into visual slides.