Loading...
Loading...
Found 26 Skills
Use when you need maximum precision on a critical task — production deployments, security-sensitive code, financial calculations, or any work where mistakes are unacceptable.
A complete software development workflow for supercharging agentic coding. Use for EVERY coding task, planning session, or when starting a new project.
Execute approved implementation plans phase by phase with automated and manual verification. Use when the user explicitly says "implement the plan", "execute the plan", or "start implementing" and has a plan file ready. Do not use for ad-hoc coding tasks without a plan.
Risk-scaled execution guardrails for repository-scoped code or configuration changes, bug fixes, debugging, refactors, code reviews, test changes, implementation plans that lead to code, and Git mutations. Skip conceptual explanations, contract-preserving prose-only edits, pure architecture design, and read-only repository status.
Build your MVP following the AGENTS.md plan. Use when the user wants to start building, implement features, or says "build my MVP", "start coding", or "implement the project".
Use when product direction, technical approach, and interface shape are known enough to implement - delivers low-uncertainty work through focused tasks, tests, checks, and review handling.
Execute approved implementation plans from scratchpad. Use when implementing a previously planned feature.
Create detailed, execution-ready implementation plans for complex or high-risk changes without coding. Use when scope is large, requirements are mostly known, and work should be broken into validated phases before execution.
Set up or log in to Karma. Use when user says "set up agent", "configure API key", "connect to Karma", "login to Karma", "log in", or before first use of any Karma skill.
[user/auto] 기존 GitHub PR body 또는 명시된 comment에 local evidence image를 업로드합니다. 명시적인 local-image 요청이나 active tk-pr-open의 정확한 evidence_required handoff에만 사용하며 generic PR·GitHub·screenshot 요청에는 적용하지 않습니다.
Systematic code analysis with evidence collection
[user/auto] 기존 repository 또는 user skill의 중복·범위·배치·trigger를 감사합니다. 기본값은 report-only이며 literal --apply 또는 current-turn approval 전에는 절대 변경하지 않습니다.