DYW
Community@catlog22 · Beijing
UCAS Electric Machine
Agent Skills by DYW
Showing 173 vetted skills indexed across 3 GitHub repositories.
team-swarm
Orchestrates ant-colony multi-agent exploration with a Python ACO controller and LLM workers.
scholar-review
Reviews academic papers pre-submission and writes structured rebuttals to reviewer feedback.
thesis-docx
Audit, revise, and format thesis Word documents with style-driven OOXML control.
maestro
Converts user intent into an auditable Session and Run execution chain.
maestro-session-manage
Inspect, complete, archive, and govern knowledge candidates for Maestro v3 sessions.
review-cycle
Routes code review workflows across session, module, and fix scopes via chain execution.
workflow-tdd-plan
Generates Red-Green-Refactor TDD task plans through a 7-phase chain workflow.
brainstorm
Orchestrates four-phase brainstorming with multi-role and single-role analysis paths.
investigate
Executes a five-phase root cause investigation chain with an Iron Law confirmation gate.
workflow-test-fix
Executes a five-phase test generation and iterative fix pipeline via chain_loader.
workflow-plan
Orchestrates a six-phase planning workflow with mode routing, conflict detection, and plan verification.
spec-generator
Generates software specifications through an 8-phase chain workflow with quality gates and auto-fix loops.
team-frontend
Orchestrates multi-role agent teams for frontend development with design intelligence and QA pipelines.
skill-iter-tune
Iteratively refine Claude Code skills through execute-evaluate-improve feedback loops.
workflow-lite-plan
Generates structured implementation plans from task descriptions via multi-angle code exploration.
workflow-lite-execute
Executes planned tasks via grouped parallel and sequential batches with optional code review.
delegation-check
Validates workflow delegation prompts against agent role definitions for content separation conflicts.
workflow-tdd-plan-plan
Generates TDD implementation plans with Red-Green-Refactor task cycles and compliance verification reports.
workflow-execute
Orchestrates autonomous execution of workflow session tasks with agent coordination and progress tracking.
team-ux-improve
Orchestrates a multi-agent pipeline to scan, diagnose, fix, and test UI/UX interaction issues.
review-code
Review code across six dimensions and generate structured severity-classified reports.
memory-manage
Updates CLAUDE.md project memory and generates API and README documentation via routed sub-commands.
ccw-chain
Orchestrates multi-step development workflows by routing user intent through chain definitions.
wf-player
Executes JSON workflow templates as DAGs with checkpoint persistence and session resume.