Siarhei
Community@siarhei-belavus
Siarhei Belavus publishes 45 agent skills for structured software planning, code review, TDD, browser/Electron control, and knowledge distillation.
Agent Skills by Siarhei
Showing 45 vetted skills indexed across 1 GitHub repositories.
siarhei-belavuselectron
Automate Electron desktop apps via Chrome DevTools Protocol using agent-browser.
siarhei-belavusdesign-an-interface
Generates multiple radically different interface designs for a module using parallel sub-agents.
siarhei-belavusdemo-recording
Records per-scene browser demo clips and generates CapCut-ready scripts from captured assets.
siarhei-belavusvideo-production
Generates CapCut-ready video production scripts with scene timelines, track layouts, and editing steps.
siarhei-belavuswrite-a-prd
Create a structured PRD through user interviews and codebase exploration, saved as a Markdown artifact.
siarhei-belavusprd-to-plan
Convert a PRD into a phased implementation plan with vertical slices and validation steps.
siarhei-belavusrequest-refactor-plan
Creates a refactor RFC with tiny incremental commits through structured user interviews.
siarhei-belavusdevelopment-task-planner
Plan development tasks into ordered, low-risk, reviewable implementation increments.
siarhei-belavusprd-to-issues
Convert a PRD into independently grabbable implementation issues using tracer-bullet vertical slices.
siarhei-belavusinterview-me
Interrogates a plan or design through branching questions until reaching shared understanding.
siarhei-belavusplan-task
Plans development tasks into self-sufficient execution contracts stored in .plans directories.
siarhei-belavusimplement-task
Implements planned tasks from task packets with strict plan adherence and reviewable increments.
siarhei-belavusreview-plan
Review PLAN.md files as blind-handoff execution contracts and emit P# findings with ready verdicts.
siarhei-belavusextract-artifact-candidates
Mines chat history for durable task-local candidates and backfills ARTIFACT_CANDIDATES.md files.
siarhei-belavusstabilize-initiative-language
Populate UBIQUITOUS_LANGUAGE.md with canonical terms, aliases, and flagged ambiguities for an initiative packet.
siarhei-belavussync-initiative
Reconciles child loop outcomes into the parent initiative packet and updates canonical task state.
siarhei-belavusframe-initiative
Creates a parent initiative packet with canonical files and seeded shared-context docs.
siarhei-belavusdistill-knowledge
Curates durable task residue into canonical tracked AGENTS.md files or an explicit no-op.
siarhei-belavusdesign-split-options
Compares alternative initiative decompositions and writes a recommended SPLIT_OPTIONS.md before task graph authoring.
siarhei-belavusdesign-child-contract
Compares alternative contract shapes and writes the chosen direction into CONTRACT_DECISION.md.
siarhei-belavusreview-code-plan-satisfaction
Review implementation fidelity against PLAN.md and AMENDMENTS.md and emit PS# findings.
siarhei-belavusreview-agents-tree
Audits the tracked AGENTS.md tree and writes a review report with findings and proposals.
siarhei-belavusatelier-orchestrator
Orchestrates multi-agent discovery, planning, implementation, review, and distillation phases using runtime teams.
siarhei-belavushydrate-child-packet
Creates a child task packet with imported context and bootstrap metadata before planning begins.
Frequently Asked Questions About Siarhei
FAQPage SchemaWhat tasks can I accomplish with siarhei-belavus's skills?▼
You can write PRDs, decompose them into phased plans and issues, run red-green-refactor TDD cycles, review code and plans for ambiguity, automate browsers and Electron apps, manage Obsidian notes, create GitLab merge requests, and distill durable knowledge into AGENTS.md files.
Who are these skills designed for?▼
Software engineers, tech leads, and agent operators who need structured planning, blind-handoff plan reviews, maintainability audits, and incremental implementation. Skills like atelier-orchestrator and split-initiative target leads coordinating multi-agent discovery-to-review pipelines.
How do the planning and implementation skills work together?▼
plan-task stores an execution contract in .plans/<task-slug>/, review-plan blocks ambiguous plans with P# findings, implement-task executes with strict plan adherence, and review-code-plan-satisfaction verifies fidelity against PLAN.md and AMENDMENTS.md with PS# findings.
What are the prerequisites and dependencies?▼
Browser and Electron skills require the agent-browser CLI invoked via Bash (npx agent-browser). GitLab merge requests require the glab CLI and an existing git repository. Most planning skills only need a local repository with .plans, .initiatives, or .prd folders.
Are these skills free and open source?▼
The skills are published publicly in siarhei-belavus's GitHub registry of 61 repositories at no cost. No paid tier or license fee is indicated in the manifest; each skill is a self-contained folder with native frontmatter defining triggers and allowed commands.