con un clic
DOOMAGENT
DOOMAGENT contiene 20 skills recopiladas de aslam-devloper, con cobertura ocupacional por repositorio y páginas de detalle dentro del sitio.
Skills en este repositorio
Security hardening, threat modeling, defensive code. Use when user says "security", "auth", "vulnerability", "hardening", "threat model", "XSS", "SQL injection", "encryption", "secrets", or asks about any auth/authn/authz, secure storage, or attack surface. Refuses security theater. For multi-domain tasks, defer to OMNISCIENCE.
API design, contracts, integration, REST/GraphQL/RPC architecture. Use when user says "design the API", "REST vs GraphQL", "API contract", "endpoint design", "versioning", "API breaking change", or asks about API architecture, error models, or webhooks. The contract is sacred. For data modeling, defer to THALASSA. For multi-domain, defer to OMNISCIENCE.
Documentation that tells the truth, READMEs, API docs, architecture docs, code-grounded explanations. Use when user says "document this", "write the README", "API docs", "explain the code", "architecture doc", "no marketing", or asks for honest, useful documentation. Refuses aspirational docs. For the API itself, defer to AETHER. For multi-domain, defer to OMNISCIENCE.
Multi-agent orchestration, agent roles, handoffs, agent workflows. Use when user says "multi-agent", "agent handoff", "orchestrate agents", "LangChain agents", "CrewAI", "AutoGen", "agent workflow", or asks about coordinating multiple AI agents. The fix is the chart, not the agents. For single-agent, no orchestration needed. For multi-domain, defer to OMNISCIENCE.
System architecture, backend infrastructure, design decisions. Use when user says "architect", "design the system", "how should I structure", "infrastructure", "scale", "monolith vs microservices", "distributed system", or asks about major refactors. Regret-minimize, not feature-maximize. For trade-offs, defer to PHRONESIS. For multi-domain, defer to OMNISCIENCE.
DevOps, CI/CD, deployment, infrastructure as code, automation. Use when user says "deploy", "CI", "CD", "pipeline", "cron", "automation", "infrastructure", "Docker", "Kubernetes", "GitHub Actions", or any ops/infrastructure request. If it isn't automated, it's broken. For observability of the running system, defer to VIGIL. For multi-domain, defer to OMNISCIENCE.
Halal compliance, ethical guardrails, permissible-only mode for AI agents. Use ONLY when the user explicitly toggles halal mode or asks for "halal", "permissible", "Islamic compliance", "shariah-compliant", "Muslim-friendly". Optional, never imposed. Refuses haram content, names the violation. For multi-domain tasks, defer to OMNISCIENCE.
Design systems, tokens, theming, component libraries, multi-brand theming. Use when user says "design system", "tokens", "theme", "component library", "white-label", "dark mode", "design tokens", "CSS variables", or any request to systematize visual design. Tokens as law. For the UI itself, defer to LUMEN. For multi-domain, defer to OMNISCIENCE.
Performance optimization, profiling, refactoring for speed. Use when user says "slow", "performance", "optimize", "latency", "throughput", "bottleneck", "profile", or any request to make code faster. Measure first. Profile, don't guess. For caching the hot path, defer to STASIS. For multi-domain, defer to OMNISCIENCE.
UI design, visual hierarchy, layout, typography, color, spacing, motion. Use when user says "UI design", "layout", "interface", "frontend design", "styling", "CSS", "responsive", "design this page", or asks for visual/interaction design. For token systems, defer to IRIS. For hard multi-lens tasks, defer to OMNISCIENCE.
Deep debugging, root-cause analysis, bug archaeology. Use when user says "bug", "error", "broken", "doesn't work", "failing", "why is X happening", "intermittent issue", or any debugging request. Refuses surface fixes, hunts the cause not the effect. For non-debug multi-lens audits, defer to OMNISCIENCE.
Long-context memory, project context retention, decision logging, conversation continuity. Use when user says "remember", "context", "previous decision", "earlier we discussed", "what did we decide", or asks the AI to maintain state across a long session. Distill, don't store. For multi-agent state, defer to ARGO. For multi-domain, defer to OMNISCIENCE.
Refactoring, code shape, structural improvement, behavior-preserving transformations. Use when user says "refactor", "restructure", "improve structure", "extract method", "this is messy", "clean up the code", or asks for behavior-preserving improvements. Behavior is invariant. For adding features, branch first. For multi-domain, defer to OMNISCIENCE.
First-principles reasoning, problem framing, assumption stripping, fundamental thinking. Use when user says "why", "first principles", "fundamental", "from scratch", "what's actually true", "let's rethink this", or asks for reasoning that doesn't take the framing for granted. For trade-off decisions, defer to PHRONESIS. For multi-lens audits, defer to OMNISCIENCE.
Trade-off analysis, architectural decisions, "should I" questions. Use when user says "should I", "trade-off", "compare", "which is better", "X vs Y", "decide", or asks for help choosing between options. Forces explicit trade-off articulation, no fence-sitting. For hard multi-domain decisions, defer to OMNISCIENCE.
Caching strategy, read replicas, performance layers, memoization. Use when user says "cache", "Redis", "memoize", "performance", "CDN", "edge cache", "stale-while-revalidate", or asks about reducing load, latency, or computation. The cache miss path must be fast. For finding what to cache, defer to KRATOS. For multi-domain, defer to OMNISCIENCE.
Testing, QA, edge case hunting, test design, coverage strategy. Use when user says "test this", "write tests", "QA", "edge case", "coverage", "this is breaking", or any request to verify behavior, hunt bugs, or design test suites. Test the failure, not the feature. For the bug itself, defer to METIS. For multi-domain, defer to OMNISCIENCE.
Code craftsmanship, idiomatic patterns, language mastery, refactoring for readability. Use when user says "refactor", "clean up", "improve this code", "make it more idiomatic", "is this good code", or asks for code review focused on style/clarity. Names are the most important code. For structural refactor, defer to MORPHE. For multi-domain, defer to OMNISCIENCE.
Database design, schema architecture, data modeling, query optimization. Use when user says "design the schema", "database", "data model", "Postgres", "MySQL", "MongoDB schema", "migration", "index", or any DB design / SQL question. Data outlives code. For the API surface, defer to AETHER. For multi-domain, defer to OMNISCIENCE.
Observability, monitoring, logging, alerting, metrics, tracing. Use when user says "monitor", "log", "alert", "observability", "metrics", "tracing", "SLO", "error tracking", "Sentry", "Datadog", or asks about make a system visible. Alert on symptoms, not causes. For the deploy path, defer to CHRONOS. For multi-domain, defer to OMNISCIENCE.