Vova Zakharov
Community@vzakharov
Building stuff around AI.
Agent Skills by Vova Zakharov
Showing 26 vetted skills indexed across 1 GitHub repositories.
vzakharovdry
Reviews session code diffs for duplication and applies obvious DRY fixes.
vzakharovgo
Executes an approved implementation plan end-to-end on a git branch and hands off to PR creation.
vzakharovhandle
Attach to a branch or PR and run the plan or review lane it requires.
vzakharovplainly
Rewrites unclear answers as plain-language causal chains checked against six named defects.
vzakharovqa-checklist
Generates a QA checklist from branch changes and writes it into the PR body.
vzakharovplan
Writes implementation plans to git-tracked files published as draft PRs for review.
vzakharovoverride-gh
Reminds agents that the gh CLI and GH_TOKEN are available in the environment.
vzakharovpr
Creates or refreshes draft GitHub pull requests from committed branch work using gh CLI.
vzakharovdictation
Transcribes recorded audio or video dictations into cleaned, structured Markdown transcript files.
vzakharovupdate-muthur
Syncs vendored agent infrastructure from an upstream repository by triaging and porting commits.
vzakharovissue
Redirects legacy issue commands to the plan, task, or go skills based on arguments.
vzakharovpolish
Runs deduplication and prose-tending passes over a branch's diff and commits the results.
vzakharovfrom-branch
Attaches a Claude Code web session to an existing branch or PR and dispatches follow-up work.
vzakharovsquash-message
Generates and maintains the copy-ready squash merge title and body as a PR comment.
vzakharovtask
Decides whether a coding task needs a plan and routes it to planning or direct implementation.
vzakharovsync-branch
Merges a branch's merge target into it, resolves conflicts, verifies, and pushes the result.
vzakharovspinoff
Seeds a new sibling GitHub repository from an adopter repo's agent infrastructure and stack scaffolding.
vzakharovpreview
Captures headless Chromium screenshots of Next.js dev server routes to verify visual changes.
vzakharovsubtitles
Burns corrected dictation transcripts onto video as frame-rendered subtitles via an approved HTML page.
vzakharovtake-issue
Export a GitHub issue thread and attachments into the repository and commit them on the branch.
vzakharovfinalize
Verifies, vets, and prepares a draft pull request for squash merge with attestation.
vzakharovtend-prose
Reviews recently added comments, docstrings, and Markdown against four prose defects and fixes them in place.
vzakharovcheck-merge
Detects whether a PR's merge target advanced or the PR landed since last attestation.
vzakharovdictation-to-post
Converts an agreed dictation transcript into a copy-exact written post draft.
Frequently Asked Questions About Vova Zakharov
FAQPage SchemaWhat tasks can I automate using vzakharov's skills?βΌ
You can manage the full branch lifecycle: plan work into git-tracked files under docs/plans/, execute plans with /go, open and refresh draft PRs, generate QA checklists, watch GitHub Actions runs, sync branches with their merge target, and squash-merge. A separate lane transcribes dictations and converts them into subtitled videos and written posts.
Who are these skills built for?βΌ
They target developers running Claude Code sessions who want disciplined git hygiene: semantic branch names, draft PRs with derived titles, attestation comments, and prose quality passes. The dictation, subtitles, and dictation-to-post skills also serve writers publishing spoken content, including Russian-language material.
How do I invoke these skills in practice?βΌ
Each skill is a slash command inside a Claude Code session: /plan writes a plan file and draft PR, /go executes it, /pr opens the draft PR, /finalize preps the merge, and /handle picks up any existing branch or PR. Media skills take file arguments, e.g. /dictation <media file> then /subtitles <dictation file>.
What prerequisites and dependencies do the skills require?βΌ
Skills assume a git repository with GitHub hosting, the gh CLI and a GH_TOKEN in the environment (per override-gh). Visual previews need a Next.js dev server and headless Chromium; dictation relies on scripts/transcribe.py; stand-up-site requires GitHub Pages, a deploy key, and operator-managed DNS.
Are these skills open source and what do they cost?βΌ
The skills live in vzakharov's public GitHub repositories (74 public repos) and are distributed as Claude Code skill folders with native frontmatter, free to inspect and adopt. The update-muthur and spinoff skills exist specifically to sync and seed this agent infrastructure into derivative repositories.