Ryan
Community@ryanportfolio
Extract Wisdom @ corewise.video
Agent Skills by Ryan
Showing 33 vetted skills indexed across 1 GitHub repositories.
ryanportfoliobro
Restates the assistant's last message in plain language without jargon.
ryanportfoliohandoff-audit
Generates a self-contained adversarial audit prompt for independent verification in a fresh session.
ryanportfolioverify-this
Verify claims by capturing baseline and treatment evidence and returning a verdict.
ryanportfoliowow-loop
Orchestrates a multi-agent implement-verify-fix loop with screenshot evidence gates for deliverables.
ryanportfoliowhy
Reviews and pressure-tests the assistant's immediately prior recommendation on explicit /why invocation.
ryanportfoliosession-hub
Coordinates parallel Claude Code sessions through a shared append-only HTML hub.
ryanportfoliomerge
Automates session-wide commit, push, pull request, and squash-merge cycles into main.
ryanportfoliolong-horizon
Orchestrates long multi-step tasks across sessions using verified executor and auditor rounds.
ryanportfoliolab
Creates throwaway single-file HTML sandboxes with live controls for tuning visual and motion parameters.
ryanportfoliopurposeful-writing
Drafts and revises prose for readers including emails, essays, proposals, and cover letters.
ryanportfoliofable-mode
Orchestrates multi-step workflows with dependent steps, load-bearing unknowns, and verification-sensitive handoffs.
ryanportfolioadopt-repo
Mirror an external repository privately and overlay firmware with privacy sweeping and project initialization.
ryanportfoliobabysit-ci
Monitors pull request checks and iterates on failures until all checks pass.
ryanportfolioaddskill
Installs new skills into a repository's .claude/skills directory with Codex adapter generation.
ryanportfoliooptimize-context
Reduces per-turn context and token load by trimming kernels, skills, and connectors.
ryanportfolioimpartial-review
Dispatches five fresh-context reviewer subagents to audit code diffs across categorized finding buckets.
ryanportfoliorecall
Retrieves and saves durable project-specific knowledge before unfamiliar edits.
ryanportfolioforge-repo-ui-skill
Synthesizes repository-specific UI design skills from existing agent skill definitions.
ryanportfoliobrainstorming
Guides product and design brainstorming with a browser-based visual mockup companion server.
ryanportfolioenhance-prompt
Rewrites rough user requests into polished, copy-paste-ready prompts for cold AI agents.
ryanportfoliowriting-plans
Generates structured implementation plans from specs before writing code.
ryanportfoliodare
Decomposes problems into first-principles assumptions and rebuilds solutions through chained subagent reviews.
ryanportfolioautomate-me
Generates a personal working-style skill from project transcripts and user preferences.
ryanportfolioarena
Spawns parallel candidate attempts at one task and merges the strongest parts into a final result.
Frequently Asked Questions About Ryan
FAQPage SchemaWhat tasks can I accomplish with Ryan's skill registry?▼
You can coordinate parallel Claude Code sessions, verify claims with fresh local evidence, babysit PR checks until green, run multi-agent perfection loops on deliverables, mirror external repos privately, and humanize prose. Skills also cover long-horizon task management, prompt enhancement, and cross-vendor code review via Codex CLI.
Who is the target user for these skills?▼
Developers running Claude Code or Codex CLI sessions who need structured verification, session coordination, and quality gates. The registry suits engineers managing multi-step work spanning context windows, reviewing code with independent agents, and writers refining prose to remove machine-generated tells before publishing.
How are these skills installed and invoked?▼
Skills are added repo-locally using the addskill skill, which handles the install contract for future Claude Code and Codex sessions. Most skills trigger via explicit dollar-sign or slash invocations like $verify-this or $babysit-ci, while some activate proactively based on context such as failing CI checks.
What prerequisites do these skills require?▼
Most skills require an active Claude Code session; impartial-review requires exposed multi-agent tools or an authenticated Codex CLI, and codex-review drives the OpenAI Codex CLI directly. Session-hub uses a shared append-only Desktop HTML hub, and adopt-repo clones external repositories under the user's account.
Do these skills support long-running or multi-session work?▼
Yes. The long-horizon skill handles work too big for one context window using verified rounds, session-hub coordinates parallel sessions through a shared HTML hub, and recall preserves durable project-specific learning. The refine skill captures friction and corrections when sessions wrap up.