AlphaLab@USTC
Community@alphalab-ustc
Alpha Lab @ USTC
Agent Skills by AlphaLab@USTC
Showing 30 vetted skills indexed across 2 GitHub repositories.
customize-system-prompt
Customize OhMyCode system prompts via OHMYCODE.md, CLAUDE.md, and system_prompt.py.
commit-conventions
Generate Conventional Commits messages with path-based Finnish scope inference for OhMyCode workflows.
add-provider
Add a new LLM provider to OhMyCode with protocol implementation and registration.
debug-ohmycode
Diagnose OhMyCode installation, provider, tool, and storage failures.
add-tool
Create OhMyCode tools with JSON Schema parameters and pytest validation.
customize-response-style
Customize terminal output styling and rendering for OhMyCode's CLI.
add-feature
Plan and implement major OhMyCode feature additions with test-first development.
using-git-worktrees
Creates isolated git worktrees for feature development while preserving the current workspace.
test-driven-development
Enforce test-first implementation with red-green-refactor cycles.
systematic-debugging
Trace technical bugs to root causes before proposing fixes.
using-superpowers
Check for applicable skills and invoke them before responding.
dispatching-parallel-agents
Dispatch independent debugging tasks to parallel agents for concurrent investigation.
executing-plans
Execute implementation plans in staged batches with review checkpoints.
finishing-a-development-branch
Verify test status and guide merge, PR, keep, or discard decisions.
brainstorming
Structures early-stage ideas into implementation-ready designs through iterative questioning and validation.
writing-plans
Create implementation plans with exact file paths and verification commands.
requesting-code-review
Request code reviews for git commit ranges before merge.
receiving-code-review
Evaluate code review feedback against the codebase before implementing changes.
writing-skills
Create, edit, and verify SKILL.md documentation with RED-GREEN-REFACTOR validation.
verification-before-completion
Require fresh command output before reporting work as complete.
subagent-driven-development
Execute implementation plans by delegating tasks to fresh subagents with spec and code quality reviews.
bench
Run OhMyCode coding benchmarks and record pass-fail outcomes with token usage.
gen-tests
Generate project-convention-compliant tests for OhMyCode modules and files.
run-tests
Run OhMyCode pytest suites and analyze failures with tracebacks.