github.com/JaviMontano/jm-adk-beta
| Skill | Added | Review |
|---|---|---|
evaluation-confidence-design skills/evaluation-confidence-design/SKILL.md Disenar evaluacion con confidence calibrada contra labeled set, stratified sampling y criterios categoricos por severidad para reducir falsos positivos sin ocultar sesgo en metricas agregadas. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
firebase skills/firebase/SKILL.md Firebase platform router: auth, hosting, functions, firestore-adjacent setup, emulators, deploy, cost. Topics: architecture, auth, cloud-functions, cost-optimization, deployment, emulator-setup, extensions, firestore-modeling, firestore-queries, firestore-security-rules, functions, hosting, scheduled-functions, setup, storage. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
frontload-prompt skills/frontload-prompt/SKILL.md Reformatea un input largo o ambiguo a estructura SPEC (Situation/Purpose/Expectations/Context) y detecta vacíos críticos antes de procesarlo. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
google-workspace skills/google-workspace/SKILL.md Router for Google APIs/Workspace work — pick one topic and load its playbook: Sheets, Docs, Slides, Drive, Calendar, Maps, Analytics/GA4, or multi-service API integration (auth/scope/quota/retry). Topics: analytics-implementation, apis, google-analytics, google-apis-integration, google-calendar-mcp, google-docs-mcp, google-drive-mcp, google-maps-integration, google-sheets-mcp, google-slides-mcp. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
guardrails skills/guardrails/SKILL.md Deterministic guard layer: pre/post tool guards, prompt filter, output contracts, secrets, constitution compliance. Thin doc over hooks/scripts. Topics: constitution-compliance, input-tolerance, integrity-chain-validation, management, output-contract-enforcer, permission-fast-path, post-tool-use-validator, pre-tool-use-guard, quality-gatekeeper, secrets-sanitization, stop-validator, user-prompt-filter. | — — — Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: fdad39c | |
hosting-infra skills/hosting-infra/SKILL.md Hosting and infrastructure: DNS, domains, SSL, CDN, serverless, backup/DR, and provider-specific deployment. Topics: backup-strategy, cdn-configuration, disaster-recovery, dns-architecture, domain-management, hostinger-deployment, infrastructure-design, serverless-patterns, ssl-management. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
iikit skills/iikit/SKILL.md Intent Integrity Kit: spec-driven development pipeline (constitution->specify->plan->checklist->testify->tasks->analyze->implement). Consumes upstream intent-integrity-chain/kit conventions. Topics: 00-constitution, 01-specify, 02-plan, 03-checklist, 04-testify, 05-tasks, 06-analyze, 07-implement, 08-taskstoissues, bugfix, clarify, core. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
integrations skills/integrations/SKILL.md Third-party service integration router: routes to one playbook for payment-integration, push-notifications, recaptcha-integration, or webhook-handling. Use when wiring an external provider (Stripe/PayPal, FCM/APNs, reCAPTCHA, inbound webhooks). | — — — Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: fdad39c | |
jarvis-bootstrap skills/jarvis-bootstrap/SKILL.md Bootstrap del Jarvis OS: crea CLAUDE.md y MEMORY.md raiz y la estructura minima de niveles N0-N4 con proyectos semilla P00/P01/P02, sin sobrescribir ediciones locales. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
jarvis-os skills/jarvis-os/SKILL.md Pack paraguas del Personal Jarvis OS: aplica COOL, detecta sector/estación, enruta a las skills del pack y materializa el método Trabajar Amplificado sobre alfa. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
kata skills/kata/SKILL.md Agentic engineering katas: proven prompt/loop/tooling patterns from JM Labs. Topics: adaptive-investigation, builtin-tool-selection, confidence-stratified-sampling, context-dilution-mitigation, critical-self-correction, custom-commands-skills, defensive-structured-extraction, deterministic-agent-loop, false-positive-criteria, fewshot-edge-calibration, headless-code-review, hierarchical-claude-memory, hub-and-spoke-isolation, human-handoff-protocol, independent-reviewer-multipass, mcp-server-configuration, mcp-structured-errors, message-batch-processing, multiagent-error-propagation, multipass-prompt-chaining, path-conditional-rules, persistent-scratchpad, plan-mode-exploration, posttooluse-normalization, prefix-caching, pretooluse-guardrails, provenance-preservation, session-resume-fork, tool-description-quality, validation-retry-feedback. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
lab-session skills/lab-session/SKILL.md Scaffold a JM Labs Lab session: the 4 canonical files (notas, hipotesis, referencias, decision) per protocol P08, missing-only and never overwriting local edits. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
legal-compliance skills/legal-compliance/SKILL.md Route legal/compliance work to one playbook: contract-review, compliance-assessment (GDPR/SOX/PCI-DSS/HIPAA/ISO 27001/NIST CSF), or compliance-framework design. | — — — Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: fdad39c | |
marketing-content skills/marketing-content/SKILL.md Marketing content production: copywriting, calendars, PR, case studies, whitepapers, video scripts, podcasts, and events. Topics: case-study-writing, content-calendar, copywriting-frameworks, event-marketing, podcast-prep, press-release, video-script, whitepaper-creation. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
market-intel skills/market-intel/SKILL.md Market and competitive intelligence: positioning, pricing, sector context, benchmarks, and partnerships. Topics: benchmarking-analysis, competitive-intelligence, competitive-positioning, market-intelligence, marketing-context, partnership-strategy, pricing-strategy, sector-intelligence. | — — — Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: fdad39c | |
mcp-engineering skills/mcp-engineering/SKILL.md Configurar MCP servers (project vs user scope, env-var expansion) y disenar contratos de error tipados con categoria y retryable. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
message-batch-orchestration skills/message-batch-orchestration/SKILL.md Orquestar Message Batches API para cargas offline con custom_id unico, fragmentacion selectiva de fallos parciales y reintento con cap. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
monthly-audit skills/monthly-audit/SKILL.md Auditoria mensual de salud del jarvis: rubrica de 6 preguntas (P22) que puntua, evidencia y prioriza acciones. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
observability skills/observability/SKILL.md Production health: monitoring, logging, alerting, health checks, and incident response. Topics: alerting-strategy, health-check-automation, incident-response, log-management, monitoring-setup. | — — — Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: fdad39c | |
official-source-verifier skills/official-source-verifier/SKILL.md Verify a decision against official sources (ADK, Agent Skills spec, GitHub/Git, framework/SDK/API/cloud docs) before changing code, docs or architecture criteria; ranks official over secondary, cites URL+access date, records the change each finding justifies, and never elevates a secondary source to authority. | — — — Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: fdad39c |