Agent Skills by Ivan Efremov
Showing 40 vetted skills indexed across 1 GitHub repositories.
gstack
Orchestrate Claude Code workflows with SKILL.md units for planning, QA, and shipping.
evaluate
Automate end-to-end smoke testing of the PMTools web app with Playwright.
generate
Generate atomic commits with npm run verify for each feature or fix.
plan-fixes
Analyze bug reports from .claude/issues/ and produce a structured work plan.
write-verify
Generate a verify-fixes.md prompt from git history and diffs.
benchmark
Automate web performance benchmarking with browse daemon and JS evaluation.
design-html
Convert design mockups into Pretext-native HTML and CSS.
design-shotgun
Generate multiple AI design variants and present them in a side-by-side comparison board.
plan-design-review
Identify missing design decisions in DESIGN PLAN documents and generate mockups.
autoplan
Run CEO, design, eng, and DX review skills sequentially with auto-decisions.
design-consultation
Generate design tokens and a DESIGN.md for product design systems.
learn
Consolidate gstack session learnings into a centralized knowledge base.
freeze
Restrict file edits to a specified directory during a session.
checkpoint
Save and resume git state, decisions, and remaining work checkpoints.
careful
Warn before destructive Bash commands via a PreToolUse hook.
cso
Audit infrastructure, dependencies, CI/CD, and AI risk with OWASP Top 10 and STRIDE.
canary
Monitor deployed apps for console errors, performance regressions, and page failures.
open-gstack-browser
Launch a visible AI-controlled Chromium browser with a live activity feed and chat sidebar.
investigate
Guide root-cause debugging through four investigation phases before proposing fixes.
document-release
Scan project diffs and update core documentation to reflect shipped changes.
health
Aggregate type checking, linting, testing, and dead-code analysis into a unified code-health score.
gstack-upgrade
Detect gstack install type and apply upgrades with migrations.
land-and-deploy
Automate PR merge, CI wait, deployment, and canary production verification.
codex
Review, challenge, or consult on code changes via the OpenAI Codex CLI.