j4flmao

Use this skill to make codebases, repositories, and documentation optimally readable and navigable by AI coding agents. Covers AGENTS.md design, repo-native instruction files, convention and constraint files, progressive context disclosure patterns, agent-optimized README structures, and workspace configuration. This skill enforces: structured metadata files, layered context loading, navigation hint systems, and machine-parseable documentation conventions. Do NOT use for: human-only documentation styling, marketing copy, or API reference generation.

Comprehensive skill for tracing reasoning paths, debugging non-deterministic agent loops, and monitoring agent behavior in production systems. Covers reasoning trace visualization, OpenTelemetry integration for agent systems, distributed tracing across multi-agent chains, decision audit logging, performance profiling, anomaly detection, cost tracking and optimization, and latency analysis for AI agent deployments.

Defines, monitors, and enforces execution-level sandboxing, performance SLA boundaries, resource limits, security isolation, network egress filters, compliance tracking, and transactional state updates. This skill enforces: resource throttling, PII scrubbers, import restrictions, network proxy compliance, atomic file locks, and circuit breakers. Do NOT use for: basic UI prompt formatting, developer code style checks, or application routing.

Use this skill to optimize and engineer prompt context windows, manage token budgets, implement dynamic context injections, handle state management, and mitigate semantic drift in LLM agent cycles. This skill enforces: structured context priority scoring, token-budget calculations, crash-resilient persistent state adapters, and drift correction pipelines. Do NOT use for: basic prompt copywriting, model evaluation datasets, or general fine-tuning prep.

Use this skill to classify agent failures, implement retry strategies with exponential backoff and jitter, design checkpoint-based state recovery, build fallback chains, manage dead letter queues, enforce error budgets, and apply chaos testing to LLM agent systems. This skill enforces: structured error taxonomies, idempotent retry logic, crash-resilient checkpoint persistence, graceful degradation cascades, and probabilistic failure injection frameworks. Do NOT use for: traditional application error handling, infrastructure monitoring/alerting, or network-level fault tolerance.

Use this skill to design and execute evaluation frameworks for LLM agents, implement trajectory testing, deploy LLM-as-judge patterns, build automated eval pipelines, and integrate agent testing into CI/CD workflows. This skill enforces: structured behavioral assertions, trajectory-vs-outcome evaluation matrices, verifier agent topologies, regression detection baselines, hallucination scoring engines, and benchmark dataset lifecycle management. Do NOT use for: unit testing traditional software, load/performance testing infrastructure, or model fine-tuning data preparation.

Use this skill to implement self-correction, reflection, human-in-the-loop (HITL), and verification layers that allow AI agents to evaluate and improve their own outputs. Covers Implement-Verify-Fix cycles, reflection patterns, HITL checkpoints, output verification, automated linting hooks, multi-stage validation, correction triggers, and quality gates. This skill enforces: structured IVF cycles, multi-layer output verification, HITL checkpoint protocols, and continuous improvement feedback mechanisms. Do NOT use for: pre-execution planning, intent classification, goal decomposition, or feedforward control mechanisms.

Use this skill to implement proactive planning and anticipation mechanisms that steer AI agents BEFORE they act. Covers OODA loops, Plan-and-Execute patterns, task decomposition, pre-flight validation, intent classification, action planning, resource pre-allocation, constraint propagation, goal decomposition trees, and anticipatory error prevention. This skill enforces: structured observation-orientation-decision-action cycles, hierarchical goal decomposition, constraint satisfaction propagation, and pre-execution validation gates. Do NOT use for: post-execution correction, feedback loops, output verification, or retrospective analysis.

Defines the rule base every skill in this pack obeys: file layout, frontmatter, vocabulary, OOP/typing/decoupling axioms, and anti-patterns. Load FIRST before authoring or applying any other skill in this pack.

Designs NestJS feature modules with clean dependency boundaries, provider scopes, and dynamic modules. Use when adding a new feature module, splitting a monolithic AppModule, or fixing tangled imports between modules.

Builds thin NestJS controllers that parse input, invoke a single use case, and map the Result to ApiResponse<T>. Use when adding HTTP endpoints, refactoring fat controllers, or applying API versioning.

Prepares a NestJS service for production: graceful shutdown, health/readiness, structured logs, config, Dockerfile, CI gates, and rollout safety. Use when promoting a service to staging/production or auditing readiness.

Designs NestJS application services as use-case classes that orchestrate domain logic and ports without leaking the framework. Use when creating new use cases, refactoring services that mix HTTP/DB/domain, or moving from anemic services to OOP.

Maps internal AppError hierarchy to HTTP responses through a single global exception filter and a Result-to-HttpException helper. Use when adding new error types, fixing inconsistent error responses, or removing scattered try/catch in controllers.

Implements the Repository pattern with TypeORM in NestJS so domain entities stay framework-free, ORM models live in infrastructure, and transactions are explicit. Use when adding persistence, refactoring leaky ORM usage, or handling transactions across aggregates.

Tests NestJS applications at unit, integration, and e2e levels using Test.createTestingModule and Supertest with hand-rolled fakes. Use when writing tests for use cases, controllers, or full HTTP flows.