rd-skills

Use this domain extension when a selected professional skill handles professional product rules for AI and ML changes involving LLMs, RAG, agents, embeddings, vector databases, classical ML model rollout, MLOps governance, model output, tool use, evaluation, hallucination risk, and AI safety. Do not use it for keyword-only, documentation-only, or UI-only mentions without domain behavior, data, risk, validation, or release impact.

Use this domain extension when a selected professional skill handles professional product rules for data warehouse, stream and batch jobs, analytics, reporting, ETL and ELT, data quality, lineage, backfill, replay, freshness, and cost-sensitive data changes. Do not use it for keyword-only, documentation-only, or UI-only mentions without domain behavior, data, risk, validation, or release impact.

Use this domain extension when a selected professional skill handles professional product rules for device, firmware, edge, protocol, hardware interaction, embedded, and IoT changes involving lifecycle, connectivity loss, rollback, compatibility, time sync, resource limits, and physical safety. Do not use it for keyword-only, documentation-only, or UI-only mentions without domain behavior, data, risk, validation, or release impact.

Use this domain extension when a selected professional skill handles professional product rules for OS, kernel, driver, native performance, networking, embedded C and C++, Rust systems, and low-level runtime changes requiring explicit memory safety, concurrency, ABI, platform behavior, syscalls, and resource cleanup. Do not use it for keyword-only, documentation-only, or UI-only mentions without domain behavior, data, risk, validation, or release impact.

Use this domain extension when a selected professional skill handles professional product rules for Android, iOS, and cross-platform mobile app changes involving offline state, push notifications, app lifecycle, permissions, privacy prompts, background execution, platform differences, and release-store constraints. Do not use it for keyword-only, documentation-only, or UI-only mentions without domain behavior, data, risk, validation, or release impact.

Use this domain extension when a selected professional skill handles professional product rules for payments, subscriptions, billing, invoices, refunds, trading, ledgers, balances, and financial workflows requiring idempotency, auditability, reconciliation, state modeling, and server-side truth. Do not use it for keyword-only, documentation-only, or UI-only mentions without domain behavior, data, risk, validation, or release impact.

Use this domain extension when a selected professional skill handles professional product rules for Web3 changes involving wallets, signatures, smart contracts, blockchain transactions, token assets, chain data, custody, private keys, Web3 auth, and on-chain/off-chain consistency. Do not use it for keyword-only, documentation-only, or UI-only mentions without domain behavior, data, risk, validation, or release impact.

Use this capability when a selected owner skill needs focused rules for verifiable completion standards and rejects vague acceptance language that cannot be tested, reviewed, or observed. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for an AI or agent changes, diagnoses, validates, reviews, deploys, or hands off code or skill work; enforces evidence-based completion, verified cause, route repair, same-pattern scans, reuse/placement rationale, and residual-risk closure. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for agent tool, command, edit, permission, sandbox, destructive-action, secret-exposure, and external-side-effect risks before execution and handoff. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for enforces explicit agent stage state, legal transitions, owner and reviewer separation, repair loops, validation freshness, and closure evidence for multi-step engineering work. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for algorithms and data structures from input scale, access pattern, time and space complexity, memory budget, streaming needs, skew, and benchmark evidence; use to avoid unbounded load-all and accidental O(n squared) implementations. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for aPI contracts with request, response, errors, authentication, pagination, idempotency, compatibility, and describable HTTP semantics. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for architecture and module rules into static checks, lint rules, dependency graph gates, type strictness, CI commands, generated-code exceptions, and migration paths. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for the simplest sufficient architecture style and requires explicit justification before service, event, or distributed complexity is added. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for produces ADR-ready architecture decisions by comparing alternatives, rejected options, decision forces, risks, mitigations, consequences, and reassessment triggers before durable architecture choices are accepted. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for durable asynchronous jobs with enqueue-after-commit, idempotency, bounded retry, timeout, status visibility, cancellation, compensation, observability, and recovery. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for identity proof from action authorization and requires object-level authorization for resource-specific operations. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for authentication token lifecycle, refresh, rotation, revocation, session fixation, MFA, cookie flags, and credential storage. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for core data changes to define backup, restore, RTO, RPO, ownership, and recovery validation before production risk is accepted. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for Make, Bazel, Gradle, Maven, npm scripts, task runners, code generation, build cache, remote execution, artifact reproducibility, or CI build graph behavior needs professional evidence. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for business rules, policies, calculations, and invariants so they are not scattered across UI, controllers, SQL, scripts, or tests. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for task-scoped Business Semantic Pack evidence across intent, vocabulary, objects, rules, workflows, memory, graph, golden cases, validation, and review without becoming a business-expert skill or runtime corpus. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for cache behavior with explicit source of truth, TTL, invalidation, stampede protection, penetration protection, consistency tolerance, and fallback. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for cI/CD pipelines with relevant build, test, lint, security checks, artifacts, deployment gates, and failure policy. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for deletion of dead code, stale feature flags, fallbacks, compatibility branches, deprecated APIs, expand-contract remnants, generated artifacts, and cleanup issues with caller search, telemetry, rollback, and owner evidence. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for cLI, TUI, and daemon interfaces with command semantics, flags, config precedence, stdout/stderr contracts, exit codes, signals, dry-run behavior, logging, idempotency, and scriptability. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for implementation readability and maintainability by checking main-flow clarity, control-flow shape, function size, nesting, naming, comments, fallback isolation, cognitive complexity, and future change navigation without duplicating placement or module-boundary design. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for low-level code element choices for variables, expressions, and statements, and bridges function, class, method, file, and directory concerns to existing structure, clarity, language, review, refactoring, and test capabilities. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for generated or modified code for correctness, maintainability, boundaries, security, tests, dependencies, naming, error handling, and hallucinated APIs. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for race, duplicate-submit, lock, optimistic concurrency, deadlock, and worker parallelism controls with deterministic conflict behavior. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for build, deploy, runtime, tenant, experiment, flag, kill-switch, rollout, mode, kind, provider, or config-driven wiring choices need typed validation, owner, rollback, observability, or cleanup. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for known and unknown consumers of API, SDK, schema, event, package, CLI, and public export changes, including compatibility, migration, deprecation, telemetry, rollout, and rollback. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for reproducible, minimal, non-root, secret-free container images with separated build and runtime concerns, health behavior, shutdown handling, provenance, and reduced supply-chain risk. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for controls ChangeForge runtime context budgets, selected references, JIT repository retrieval, tool-output boundaries, compaction snapshots, branch summaries, route repair, and overhead evidence for routing decisions. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for relevant AI coding context with requirements, architecture, contracts, files, constraints, tests, and quality gates while preventing context drift. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for aPI, schema, event, and external contract changes to protect consumer expectations and compatibility through executable contract evidence. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for controller responsibilities for transport parsing, validation handoff, auth context, response mapping, and error boundaries without core business logic. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for C/C++ code, native extensions, embedded modules, ABI/FFI, parsers, atomics, CMake/toolchain, or performance-sensitive native work needs RAII, undefined-behavior, ownership, sanitizer, thread-safety, portability, resource-safety, graph/memory freshness, or validation-evidence review. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.

Use this capability when a selected owner skill needs focused rules for JSON, YAML, XML, CSV, Protobuf, Avro, Parquet, OpenAPI schema, serialization, parser behavior, compatibility, coercion, field evolution, or wire-format contracts need professional evidence. Do not use it as a standalone owner for broader implementation, review, release, or documentation work.