pgoell
UserCollection of my personal claude skills
Categories
Indexed Skills (39)
capturing-session-learnings
Capture session learnings into AGENTS.md or CLAUDE.md. Triggers on /revise-agents-md and /revise-claude-md.
improving-instructions
Audit and improve AGENTS.md and CLAUDE.md files across project and user scopes.
databricks-core
Databricks CLI operations: auth, profiles, data exploration, and bundles. Contains up-to-date guidelines for Databricks-related CLI tasks.
crafting-presentations
Use when the user wants to build a multi-slide HTML presentation with slide-type templates, a deck-stage engine, and an optional two-window presenter view for live sharing in Teams, Zoom, or Meet. For one-off single-file slide artifacts, see workbench:crafting-html. For brand theming, see workbench:crafting-design-systems.
writing
Draft, review, and finish long form prose, essays, talks, newsletters, memos, and briefings.
presentations
Use when the user wants to design the content of a slide presentation, from audience brief through critiqued storyboard. Produces a markdown deck.md with structured per-slide front-matter. For a written prose talk instead of a deck, use writing:writing with format talk. For rendering a deck.md to HTML, use workbench:crafting-presentations.
emil-design-eng
This skill encodes Emil Kowalski's philosophy on UI polish, component design, animation decisions, and the invisible details that make software feel great.
quizzing-a-topic
Use when the user wants to be taught and quizzed on a topic, theme, or concept they name (not the current session), to verify they deeply understand it. For example "teach me about Kafka consumer groups and test me" or "quiz me on this repo's auth flow".
quizzing-the-session
Use when the user wants to be taught and quizzed on the work from the current session, what was just built, decided, debugged, or changed, to verify they deeply understand it. Not for updating agent instruction files; that is capturing-session-learnings.
copilot
Use when the user wants semi-autonomous, human-in-the-loop feature shipping where the user drives the brainstorm and spec by hand and the agent (the copilot) automates the plan, implementation, docs, audit, and PR using .workbench/autopilot.md. The interactive sibling of workbench:autopilot.
exporting-decks-to-pptx
Use when the user wants to export or convert an existing workbench HTML deck into a native, editable PowerPoint file (.pptx) with real text boxes, autoshapes, embedded images, and speaker notes. Triggers on requests like "make a PowerPoint version of this deck" or "export the deck to pptx". For building a new HTML presentation, see workbench:crafting-presentations.
perfecting-presentations
Use when the user wants to review, audit, polish, or perfect an existing HTML slide deck until it is genuinely done, for example "review this deck until it is perfect", "make this presentation crisp", "audit my slides". Runs a strict convergence loop, deterministic hard gates plus fresh-context judge panels with adversarial verification, terminating only when consecutive review rounds come back dry. For building a deck see workbench:crafting-presentations; for designing slide content see writing:presentations.
creating-skills
Use when scaffolding, iterating on, pressure-testing, optimizing the description of, or extracting from a session a Claude Code or Codex skill in any plugin marketplace. Detects the marketplace shape automatically (Claude Code, Codex, or both) and adapts to the repo's own conventions for paths, author, license, test layout, and version-tracking. Triggers on phrases like "create a skill", "scaffold a skill", "iterate on this skill", "pressure-test", "optimize the description", or "turn this conversation into a skill".
jira
Search Jira, manage issues, add comments, transition work, and inspect sprints.
frontend-design
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.
calendar
View agendas, manage events, check availability, and update calendars with gws.
gmail
Search, read, send, and manage Gmail messages, drafts, labels, and filters with gws.
playground
Creates interactive HTML playgrounds, self-contained single-file explorers that let users configure something visually through controls, see a live preview, and copy out a prompt. Use when the user asks to make a playground, explorer, or interactive tool for a topic.
research
Research complex topics with sources, synthesis, review, and a final report.
claude-codex-bridge
Align Claude Code and Codex project files, settings, hooks, agents, and plugins.
tmux
Use when the user wants to control interactive terminal programs through tmux, drive REPLs, debuggers, TUI tools, or experiment with steering nested Claude Code or Codex CLI sessions by sending keys and reading pane output.
autopilot
Autopilot for autonomous feature shipping from brainstorm to PR using .workbench/autopilot.md.
brainstorming
Use before creative work to grill the user toward a shared design concept through exhaustive sequential question and answer.
crafting-design-systems
Use when the user wants to create or edit a workbench design system, a directory of CSS variable overrides, optional components, and optional images that themes HTML output from workbench producer skills.
crafting-html
Use when producing a standalone HTML artifact that is not already covered by another workbench skill (writing-spec, writing-plans, brainstorming, systematic-debugging) or by research:research. Covers PR walkthroughs, code explainers, slide decks, status and incident reports, design prototypes, SVG illustrations, custom editing interfaces, and similar single-file HTML outputs. Bundles 20 reference examples plus an index file (21 files total) for inspiration; read individual files lazily, not all at once.
dispatching-parallel-agents
Use when facing two or more independent tasks that can be worked on concurrently without shared state or sequential dependencies.
subagent-driven-development
Use when executing an implementation plan with fresh subagents, task ownership, and review gates in the current session.
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
terse-mode
Use only when the user explicitly asks to switch the session into terse mode, terse-mode, less tokens mode, token saving mode, or /terse-mode.
test-driven-development
Use when implementing any feature, bugfix, refactor, or behavior change before writing implementation code. Enforces test-first RED-GREEN-REFACTOR discipline.
using-workbench
Load Workbench skill rules and choose the right Workbench skill.
verification-before-completion
Use when about to claim work is complete, fixed, passing, ready, or safe to merge, especially before commits, pushes, PRs, or handoff summaries.
visualizing-options
Use when a design discussion needs visual options shown in a browser: mockups, layout comparisons, wireframes, or architecture diagrams that the user can click to choose between. Spawns a local server and renders content fragments.
writing-plans
Use when a design spec or approved requirements need to become a concrete, step-by-step implementation plan before code changes.
writing-spec
Use when a design discussion is ready to be written into a spec document, before any implementation planning. Synthesizes the conversation into a spec, runs a fresh-eyes self-review subagent, and gates on user approval before handing off to writing-plans.
pyramid
Structure memos, recommendations, briefings, and decision documents with the Pyramid Principle.
tech-doc
Draft, review, and finish tutorials, how to guides, references, and explanations.
confluence
Search, read, create, and update Confluence pages.
databricks-docs
Use when the user asks a Databricks product question (Apps, DABs, Jobs, Lakebase, Model Serving, Pipelines, Unity Catalog, SQL Warehouses, MLflow, serverless, secrets, workflows) that databricks-core's CLI cheatsheet does not answer. Resolves a canonical URL from references/docs-index.md and fetches it with the host agent's WebFetch tool.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.