Agent Skills by Martin Hylleberg
Showing 23 vetted skills indexed across 1 GitHub repositories.
workflow-guide
Select workflow modes by assessing task scope, stakes, and parallelism.
adr
Document architectural decisions using a standardized ADR template with context, options, decisions, and consequences.
implement-phase
Orchestrate an implementation plan phase by delegating to subagents and producing a completion report.
brainstorm
Guide rough ideas into actionable plans through Socratic questioning and structured analysis.
eval-harness
Run configurable capability and regression evaluations with multiple grader backends.
context-saver
Save session context to a standardized CONTEXT-*.md file.
codebase-research
Decompose codebase queries into parallel sub-agent research tasks and synthesize structured reports.
create-plan
Create executable implementation plans with phases, tasks, and ADRs.
code-review
Review code changes for architectural conformance and quality gates.
iterate-plan
Refine implementation plans by updating phases and syncing Task tools.
agent-creator
Generate composable AI agent systems with DAG-based planning in NestJS.
skill-visualizer
Generate interactive D3.js HTML visualizations of skills and dependencies.
team-implement-plan
Coordinate a small team to execute plans with role-based tasks and phase reviews.
e2e-testing
Automates end-to-end web testing with Playwright MCP and regime-driven YAML workflows.
continuous-learning
Extract and persist learning patterns from Claude Code sessions.
verification-loop
Automates Build, Types, Lint, **Tests, Security, and Diff checks for projects.
team-implement-plan-full
Coordinates parallel implementation waves with a central reviewer and lead for shared-file governance.
team-create-plan
Orchestrate Architect, Risk Analyst, and Researcher roles to generate ADR-backed implementation plans.
implement-plan
Coordinate multi-phase implementation plans and delegate each phase to sub-skills.
prompt-generator
Generates phase-based orchestration prompts from templates and saves them to docs/prompts.
security-review
Review code changes against OWASP Top 10 and generate remediation guidance.
strategic-compact
Suggest context compaction at logical coding boundaries using Claude Code hooks.
team-brainstorm
Orchestrate adversarial multi-perspective brainstorming with agent teams to analyze ideas.