AlexanderNarbaev
Community@AlexanderNarbaev · Russia, Moscow
AlexanderNarbaev maintains 43 orchestration skills for opencode-swarm, covering multi-agent code review, test execution, planning, and repository governance protocols.
Agent Skills by AlexanderNarbaev
Showing 43 vetted skills indexed across 1 GitHub repositories.
AlexanderNarbaevrunning-tests
Executes Bun test suites safely with scope guards, per-file isolation loops, and failure classification.
AlexanderNarbaevdesign-docs
Generates and syncs language-agnostic design documentation with stable section IDs and traceability.
AlexanderNarbaevdeep-dive
Audits codebase scope with parallel explorer waves, dual reviewers, and critic challenges.
AlexanderNarbaevdiscover
Maps repository structure, API surfaces, and governance rules through read-only exploration.
AlexanderNarbaevcouncil
Orchestrates a three-agent advisory council with web research, parallel dispatch, and synthesis.
AlexanderNarbaevbrainstorm
Guides structured discovery dialogue through seven phases to produce a reviewed specification.
AlexanderNarbaevplan
Generates structured implementation plans with spec traceability, clarification funnels, and QA gate persistence.
AlexanderNarbaevissue-ingest
Ingests GitHub issues, localizes root causes, and generates resolution specs.
AlexanderNarbaevclarify
Runs a four-stage clarification funnel with critic review before surfacing user decisions.
AlexanderNarbaevresume
Resumes an approved project plan from current state with swarm identity reconciliation.
AlexanderNarbaevdeep-research
Orchestrates multi-round web research with parallel synthesis, claim verification, and cited reports.
AlexanderNarbaevcoprocessor
Enforces dual-process reasoning, memory journaling, and structured output protocols for AI agents.
AlexanderNarbaevengineering-conventions
Enforces engineering invariants for plugin initialization, subprocesses, and test isolation in opencode-swarm.
AlexanderNarbaevcodebase-review-swarm
Runs quote-grounded multi-track codebase audits with coverage closure and critic validation.
AlexanderNarbaevcritic-gate
Enforces critic review of plans with revision loops before execution begins.
AlexanderNarbaevswarm-pr-feedback
Ingest, verify, and resolve pull request feedback with a tracked closure ledger.
AlexanderNarbaevclarify-spec
Resolves specification clarification markers through structured one-at-a-time user questions.
AlexanderNarbaevswarm-pr-review
Runs multi-lane parallel pull request review with independent validation and critic challenge.
AlexanderNarbaevcommit-pr
Enforces a structured commit, validation, and pull request publication protocol for GitHub repositories.
AlexanderNarbaevexecute
Orchestrates task execution with QA gates, coder retries, and completion evidence.
AlexanderNarbaevwriting-tests
Guides writing bun:test unit tests with mock isolation and cross-platform CI rules.
AlexanderNarbaevconsult
Answers advisory questions by consulting domain SME agents with bounded evidence and cached guidance.
AlexanderNarbaevphase-wrap
Executes phase-boundary closeout with retrospective evidence, drift gates, and council reviews.
AlexanderNarbaevspecify
Generates structured feature specifications with QA gate selection and clarification funnels.
Frequently Asked Questions About AlexanderNarbaev
FAQPage SchemaWhat tasks can I accomplish with AlexanderNarbaev's opencode-swarm skills?▼
You can run structured development modes: deep PR reviews, full-repo security/QA/performance audits, safe test execution with bun, TDD cycles, bug diagnosis loops, GitHub issue ingestion, spec and plan creation, merge conflict resolution, and multi-round council research with cited reports.
Who are these skills designed for?▼
They target engineers and architects working inside the opencode-swarm repository or using its architect-driven mode system. Maintainers handling PR feedback, CI failures, release notes, and large multi-session workstreams benefit most, as do reviewers needing low-false-positive, quote-grounded audit reports.
How are the skills loaded and executed at runtime?▼
Skills load on demand when the architect emits a mode signal such as [MODE: DEEP_DIVE] or [MODE: EXECUTE]. Each skill is a full execution protocol defining dispatch, reviewer/critic validation, QA gates, and artifact writes under .swarm/ directories, with some like swarm-pr-review disabling direct model invocation.
Are these skills open source and what do they cost?▼
The codebase-review-swarm skill is explicitly MIT licensed; other skills carry no stated license or fee. They are prompt-protocol skills distributed via the repository, so usage cost is limited to the model inference and compute of the host opencode-swarm environment.
What prerequisites or dependencies do the skills require?▼
Skills assume the opencode-swarm environment with its architect command layer, .swarm storage, and plugin system. Testing skills require bun and bun:test with cross-platform CI; review skills need git history, GitHub issues/PRs, and adherence to AGENTS.md and docs/engineering-invariants.md conventions.