Andrii Shafar
Community@reidond · Ukraine
Agent Skills by Andrii Shafar
Showing 26 vetted skills indexed across 1 GitHub repositories.
ai-improvement-tracker
Records testable improvement hypotheses for AI infrastructure changes after workflow updates.
skill-reviewer
Audit Claude Code skills for specification compliance and workflow placement.
parallel-deep-research
Coordinate multi-agent research with deduplicated source ownership and cited report generation.
research-source-planner
Build a deduplicated source manifest with URL normalization and ownership assignment.
prompt-domain-reviewer
Review AI prompts for domain accuracy and methodology gaps.
review-prompts
Review AI prompt templates for structure, domain alignment, and LLM best practices.
fable-orchestrate
Decompose engineering goals into delegated subtasks and coordinate verified results.
product-brief
Create structured product briefs with UX copy and implementation notes.
agent-delegate
Delegate scoped engineering tasks to specialized worker agents with structured briefs and verification.
skill-content-reviewer
Review Claude Code skill content for accuracy, depth, and practical guidance.
spec-driven-dev
Guide software teams through structured specification and planning workflows before implementation.
ai-changelog
Record AI infrastructure changes in a structured changelog format.
learning-consolidator
Consolidate project learning logs into AI rules, skills, and conventions.
branch-switch
Switch Git branches while preserving uncommitted changes via stashes.
session-retrospective
Analyze Claude Code session history to identify recurring agent behavior problems.
prompt-eng-reviewer
Review AI prompt templates for quality issues and output reliability.
skill-researcher
Investigate topics to produce structured research briefs for skill creation workflows.
skill-creator
Create structured Claude Code skills with YAML frontmatter and validation workflows.
ai-docs-lookup
Retrieve current official AI provider documentation before answering technical questions.
task-learnings
Record reusable project knowledge after completed software development tasks.
post-task-review
Reviews completed dev tasks for code quality issues and documentation gaps.
prompt-engineering-conventions
Apply structured conventions to create, review, and maintain system prompts.
commit-message
Generate conventional commit messages from staged diff analysis.
plan-critic
Review implementation plans for security gaps, missing requirements, and resilience issues.