Tengwei Cai
Community@caitengwei · Hangzhou, China
Building model inference and RayServe platform @alipay
Agent Skills by Tengwei Cai
Showing 22 vetted skills indexed across 1 GitHub repositories.
caitengweihumanizer-zh
Rewrite Chinese text to remove detectable AI writing patterns.
caitengweicodebase-analyzer
Convert an unknown repository into structured documentation and engineering knowledge.
caitengweisync-upstream-skills
Synchronize upstream skills repositories into a local workspace using git subrepo pulls.
caitengweidefuddle
Extract clean markdown content from web pages by removing clutter.
caitengweiobsidian-markdown
Write Obsidian Flavored Markdown with wikilinks, embeds, callouts, and frontmatter.
caitengweiobsidian-bases
Create Obsidian .base files with filters, formulas, and views.
caitengweiobsidian-cli
Manage Obsidian vault notes, tasks, and properties from the command line.
caitengweijson-canvas
Create and edit JSON Canvas diagrams for Obsidian with valid nodes and edges.
caitengweiusing-git-worktrees
Validate Git worktree directory safety and tracked state before feature work.
caitengweitest-driven-development
Guide test-first development through a failing test cycle with Red-Green-Refactor.
caitengweisystematic-debugging
Enforce a four-phase root-cause investigation workflow before fixing bugs.
caitengweiusing-superpowers
Enforce pre-response skill invocation before replying to any conversation turn.
caitengweidispatching-parallel-agents
Dispatch debugging tasks to parallel agents for independent failure domains.
caitengweiexecuting-plans
Execute written implementation plans with verification checkpoints and task tracking.
caitengweifinishing-a-development-branch
Verify tests and choose one of four Git branch completion actions.
caitengweibrainstorming
Guides product or engineering ideas to an approved design specification.
caitengweiwriting-plans
Convert feature specifications into test-driven implementation plans with file paths and execution steps.
caitengweirequesting-code-review
Triggers Git-based code review using BASE_SHA and HEAD_SHA ranges.
caitengweireceiving-code-review
Inventory unresolved threads, verify requirements against the codebase, and classify as fixed declined or blocked.
caitengweiwriting-skills
Apply test-driven development to create compliant agent Skill documentation.
caitengweiverification-before-completion
Require fresh command execution and evidence before reporting completion status.
caitengweisubagent-driven-development
Dispatch isolated subagents per task and validate through sequential spec and code reviews.