Agent Skills by tensei
Showing 89 vetted skills indexed across 3 GitHub repositories.
cocos-creator-3x-cn
Guide Cocos Creator 3.x TypeScript game development with component systems and best practices.
cocos-creator-2x-cn
Codify Cocos Creator 2.x development standards for TypeScript and component architecture.
adversarial-review
Orchestrate adversarial code reviews with opposing-model reviewers and synthesized verdicts.
cocos-mcp-setup
Install and configure the cocos-mcp-server MCP extension in Cocos Creator projects.
omo
Route code-analysis tasks to specialized agents based on task type and risk.
using-git-worktrees
Create isolated Git worktrees with safety checks and directory selection.
test-cases
Convert product requirements into traceable QA-ready Markdown test cases.
test-driven-development
Enforce test-driven development with failing tests before production code.
reverse-spec-from-repo
Reverse-engineer a prompt-spec suite from a codebase into spec/ with evidence.
systematic-debugging
Enforce a four-phase root-cause debugging workflow before proposing fixes.
gemini
Execute Gemini AI code analysis and generation tasks via CLI.
codex-skill
Orchestrate Codex/GPT execution to turn high-level requests into working code.
youtube-transcribe-skill
Extract YouTube video subtitles into timestamped text transcript files.
do
Coordinate seven-phase feature development tasks across multiple automation agents.
spec-kit-skill
Coordinates constitution-based feature development with GitHub Spec-Kit across seven workflow phases.
root-cause-analyzer
Guide users through the 5 Whys method and document findings in a knowledge base.
architecture-validator
Validate plan.md completeness and alignment with feature specifications.
code-execution-helper
Execute parallel batch scans with scanFile and return token-efficient summaries.
using-superpowers
Invoke the Skill tool before responding to verify applicable skills.
dispatching-parallel-agents
Dispatch parallel agents to investigate independent failures across files or subsystems.
product-requirements
Translate ambiguous project requirements into testable PRDs saved to docs/{feature-name}-prd.md.
kb-create
Create database entries with interactive prompts and automatic quality validation scoring.
codex
Delegate complex code analysis and refactoring tasks to Codex AI.
deep-research
Orchestrate multi-agent deep research into a structured final report.