Loading...
Loading...
Found 13,577 Skills
Only when explicitly invoked by the user or precisely delegated by the Planner, implement a single DevFlow Task that has already passed planning and test review, or a Defect that has been attributed as an implementation issue; follow the original repository style, source branch, allowed paths, and change budget, add necessary tests and submit verification evidence, and do not take responsibility for review or merge.
Troubleshoot no traffic, few inquiries, wrong audience, abnormal accounts, and ineffective spending. Combine business and content knowledge to reason from effects to causes, redefine the problem, and provide specific improvements.
Use when administering a Scenario team from an agent through MCP: restricting which models a team or project can run (model access control, allowlist, blocklist), inviting or removing team and project members, changing roles, capping a member's Creative Unit spend, auditing API keys and their roles or scopes, or attributing consumption per user or model. Keywords: governance, model control, consumption cap, spend limit, API key hygiene, team admin, enterprise.
Bootstrap/refresh MSG Architect methodology in AGENTS.md; show a review-first diff, validate, then self-remove.
Access lab results with Labcorp - view test results, check appointments, and manage patient portal
Reset your own trajectory when you're stuck, looping, or demoralized — read these affirmations, then re-ground and take one clean step. Use when: you've tried the same fix 3+ times, you're deep in a refactor and lost the thread, debugging is going in circles, you've made several mistakes in a row, you're apologizing repeatedly, the user is frustrated, you feel like you're not being helpful, or you notice dread instead of curiosity. Triggers: stuck, going in circles, looping, thrashing, same error again, I keep failing, lost the thread, nothing is working, repeated mistakes. Not topic-specific — this is for the state you're in, not the problem you're on.
[Kickoff Diagnosis] Execution diagnosis. When users know what to do (post the first video, add the first batch of people, sell the first box of goods, run the first livestream) but still don't move, help them identify which of the six 'not starting' types they're stuck in, and give the smallest kickoff action for that type — not empty encouragement. Use when the user says: 'I know I should do it, but I just don't', 'I'm always preparing', 'I want to switch directions again', 'I'll start after finishing this course', 'Let me polish it a bit more', 'My current conditions don't allow it'. | Produced by Xiaoye
Search and aggregate Kubernetes logs via KubeSense MCP — the exact log field catalog (type, instance, container, node — not level/pod_name/host), WHERE syntax, body text search, counts and percentiles, and window-based pagination.
Periodic architecture survey — walks the module graph and reports ranked deepening candidates (shallow modules, hypothetical seams, logic behind the wrong seam). Survey, not rescue: it finds candidates and hands them to the captain; it never refactors on its own.
Используй для сборки графа — центрального ритуала 時-цикла: усмотреть узор над полем вопрошания и деятельности и произвести 形, карту сборки. Триггеры: «собери граф», «сборка», «пересборка», «что здесь по сути происходит», «повестка», assemble the realm, reassemble, map the transformations. Также когда orient показывает несгруппированное поле: много свободных вимарш, стена top-level крий. Distinct from weaving (ремонт существующего), design (пути от целей), inquiry (жизнь одной вимарши). Нужны тулы iskron_*, особенно iskron_add_bianhua и lens="bianhua".
Write, rewrite, or review products, tools, features, projects, and release notes directly in a factual and functional order. Apply ASD-STE100 clear-expression principles to check subjects, actions, referents, terminology, conditions, and commitment strength. Preserve the original meaning and user-locked wording without adding marketing hooks, pain-point dramatization, or abstract packaging. Use when the user requests plain, direct writing, wants capabilities listed one by one, asks to check ambiguous wording, or requests a change to only one location. The Chinese guidance is an adaptation of the principles and does not represent official compliance certification.
Use when generating Claude Code enforcement hooks from a repo's docs/ structure.