ryanportfolio

Use when writing or editing any prose a person will read (site copy, guides, docs, READMEs, UI text, titles, captions, emails, PR descriptions, release notes) or when a draft reads as machine-made.

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

Use when implementing any feature or bugfix, before writing implementation code

Use when creating new skills, editing existing skills, or verifying skills work before deployment

Use when the user asks to reduce per-turn context or token load, trim kernels, skills, or connectors, or propagate a generic context optimization to the starter.

Use when brainstorming or designing a product, interface, workflow, architecture, or behavior change with unresolved goals or material tradeoffs; not for routine or fully specified work.

Use when the user asks to humanize, de-AI, de-slop, voice-match, or review prose for AI tells before publishing.

Use when the user explicitly asks to lab or prototype a visual, UI, motion, or game-feel element with live tuning before production implementation.

Add a new skill to this repo so it's available in every Claude Code web session. Use when the user says $addskill, asks to "add a skill", "install a skill", "create a skill", or wants a third-party skill (e.g. superpowers) to show up in the available-skills…

Use when the user asks to brainstorm, explore, compare approaches, or design a feature, or when a requested product, interface, workflow, architecture, or behavior change has unresolved goals, competing directions, material tradeoffs, or high-cost ambiguity.…

Research current GitHub agent skills for frontend design, UI, UX, web, accessibility, motion, and creative implementation; determine which guidance is actually relevant to the active repository; and synthesize one lean repo-local design skill. Use when a user…

Research current GitHub agent skills for frontend design, UI, UX, web, accessibility, motion, and creative implementation; determine which guidance is actually relevant to the active repository; and synthesize one lean repo-local design skill. Use when a user…

Devil's advocate on change just made — argue honest case AGAINST before it lands. "Any reason not to?" Deliberate second-thoughts pass on diff in front of you. Surfaces scope creep, blast radius, irreversibility, simpler path skipped, YAGNI, "doesn't belong…

Code-quality and performance checklist for this project's stack. Use BEFORE making non-trivial changes — new features, refactors, perf work, bug fixes — to spot common pitfalls (N+1 queries, duplicate query-cache keys, non-passive scroll listeners, missing…

Use when a git merge/rebase/cherry-pick is in progress and has left unmerged paths or conflict markers (<<<<<<< / ======= / >>>>>>>) — especially when two parallel sessions edited the same file (the same page/component) and one side now conflicts. Reviews…

Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy.

Always-on AI-tell stripper: apply its pattern check to everything written for humans (chat prose, commits, PR bodies, docs, UI text). Also use when the user says /unslop, "unslop this", or points at text or a file to clean.

Spawn N parallel candidate attempts at one task, pick the strongest as base, graft the losers' best parts in. Use when the user says $arena, "arena this", or when one attempt at a non-trivial artifact would lock in the wrong shape.

Use for "automate me", "$automate-me", "create/update my -mode skill", or "turn my preferences / working style into a skill". Mines the current project's transcripts plus direct questions, then drafts a personal <handle>-mode skill.

Watch a PR's checks and iterate on failures until green. Use for $babysit-ci, "watch CI", "fix CI", "get the checks green", or when a PR is waiting on failing or pending checks.

Restate the assistant's last message in plain human language, no jargon. Use when the user says $bro, "in plain english", "dumb it down", or "what does that actually mean".

Always-on AI-tell stripper: apply its pattern check to everything written for humans (chat prose, commits, PR bodies, docs, UI text). Also use when the user says $unslop, "unslop this", or points at text or a file to clean.

Verify a claim with fresh local evidence: restate it falsifiably, capture baseline and treatment, compare, return VERIFIED, NOT VERIFIED, or INCONCLUSIVE. Use for $verify-this, "prove it works", "did this fix it", "show me the evidence".

Use when the user asks to review, audit, or stress-test recent code changes with fresh independent agents; requires exposed multi-agent tools or an authenticated Codex CLI.

Add a new skill to this repo so it's available in every Claude Code web session. Use when the user says $addskill, asks to "add a skill", "install a skill", "create a skill", or wants a third-party skill (e.g. superpowers, impeccable) to show up in the…

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy.

Rewrite a rough or under-specified request into a polished, copy/paste-ready prompt that any agent or LLM can act on cold. Use when the user gives a quick description of something they want done and asks for it to be turned into a better prompt — e.g., "make…

Generate a self-contained, copy/paste-ready audit prompt so a SEPARATE fresh session can independently verify a change made in the current session. Use when the user says $handoff-audit, or asks to "double-check this in another session", "have another session…

Turn on Auto-Merge Mode for the session — going forward, every completed task is automatically committed, pushed, opened as a PR, and merged into main (with conflict resolution), no per-merge confirmation. Use when the user says $merge, "auto-merge for the…

Reduce per-turn context/token cost (the always-loaded weight: CLAUDE.md kernel, MEMORY.md index, the deferred MCP tool-name list, MCP instruction blocks, and the injected available-skills list) — then propagate the portable, generic wins up to the…

Use for work too big for one context window: long multi-step tasks, progress lost to compaction or failed retries, work spanning hours or sessions, or when the user says $long-horizon or asks to run a task in verified rounds.

Verify a claim with fresh local evidence: restate it falsifiably, capture baseline and treatment, compare, return VERIFIED, NOT VERIFIED, or INCONCLUSIVE. Use for /verify-this, "prove it works", "did this fix it", "show me the evidence".

Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy.

Use when the user asks to humanize, de-AI, de-slop, voice-match, or review prose for AI tells before publishing.

Use when creating new skills, editing existing skills, or verifying skills work before deployment

Cross-vendor second-opinion review. Drives OpenAI Codex CLI (codex exec review, gpt-5.6-sol, high reasoning) over a PR, branch, commit, or uncommitted diff, then verifies each finding. Trigger: $codex-review, "have Codex/Sol review this".

Multi-agent perfection loop for any deliverable. Recon, one spec, one implementer, adversarial screenshot-verified critique until an evidence gate passes. Use when the user says $wow-loop, asks for "wow factor" or "dial it to 11".

Use whenever a repo-local skill is being added, installed, or created for future Claude Code and Codex sessions, whether the user asked or you decided to write one; writing-skills covers authoring, not this repo's install contract.

Use when assessing Truenote security posture, reviewing FedRAMP alignment, checking P0/P1 or later phase progress, deciding what security work comes next, validating security claims, or refreshing the public security capabilities brief.

Use when assessing Truenote security posture, reviewing FedRAMP alignment, checking P0/P1 or later phase progress, deciding what security work comes next, validating security claims, or refreshing the public security capabilities brief.

Use when the user asks to brainstorm, explore, compare approaches, or design a feature, or when a requested product, interface, workflow, architecture, or behavior change has unresolved goals, competing directions, material tradeoffs, or high-cost ambiguity.…

Design, implement, review, or polish Truenote's React frontend. Use for any user-facing change under artifacts/rag-app involving pages, components, layout, visual hierarchy, UX copy, forms, tables, navigation, responsiveness, motion, accessibility,…

Design, implement, review, or polish Truenote's React frontend. Use for any user-facing change under artifacts/rag-app involving pages, components, layout, visual hierarchy, UX copy, forms, tables, navigation, responsiveness, motion, accessibility,…

Turn on Auto-Merge Mode for the session — immediately integrate any already-completed verified work, then automatically commit, push, open a PR, and merge every future completed task into main (with conflict resolution), no per-merge confirmation. Use when…

Rewrite a rough or under-specified request into a polished, copy/paste-ready prompt that any agent or LLM can act on cold. Use when the user gives a quick description of something they want done and asks for it to be turned into a better prompt — e.g., "make…

Generate a self-contained, copy/paste-ready audit prompt so a SEPARATE fresh session can independently verify a change made in the current session. Use when the user says $handoff-audit, or asks to "double-check this in another session", "have another session…

Use when designing, reviewing, or implementing interactive charts, diagnostics, evidence plots, or chart-centered question answering in an insurance actuarial workflow.

Use when designing, refining, reviewing, or implementing Prediction Lab or insurance actuarial modeling UX, especially AI-agent experiments, evidence, versioning, collaboration, and model review.

Use when comparing, reviewing, reconciling, adjudicating, or finalizing multiple completed actuarial UX synthesis passes, especially Prediction Lab research syntheses, before using findings for product or design decisions.

Use when comparing, reviewing, reconciling, adjudicating, or finalizing multiple completed actuarial UX synthesis passes, especially Prediction Lab research syntheses, before using findings for product or design decisions.

Use when independently researching Prediction Lab, incumbent insurance actuarial modeling software, legacy pricing workflows, practitioner friction, workarounds, latent UX problems, or when synthesizing multiple independent research passes.

Use when independently researching Prediction Lab, incumbent insurance actuarial modeling software, legacy pricing workflows, practitioner friction, workarounds, latent UX problems, or when synthesizing multiple independent research passes.

Use when designing, reviewing, or implementing interactive charts, diagnostics, evidence plots, or chart-centered question answering in an insurance actuarial workflow.

Use when designing, refining, reviewing, or implementing Prediction Lab or insurance actuarial modeling UX, especially AI-agent experiments, evidence, versioning, collaboration, and model review.

Use when building or refining a web prototype from a live reference site's design, layout, responsive behavior, interactions, motion, or page transitions; when capture tooling is available or the user says /reference-site-prototyping.

Produce a copy/paste-ready prompt that starts a fresh session on the next round of a multi-round effort. Use when the user says $next-round, asks for a handoff prompt, or just finished a round and wants the next session to continue.

Use when adding or creating any DSH (DeepSeek Harness) plugin, panel, tool, or UI feature. Author plugins the permanent profile-bundle way, never the temporary dynamic way (cordis_define/cordis_run), which dies on restart.

Use when building or refining a web prototype from a live reference site's design, layout, responsive behavior, interactions, motion, or page transitions; when capture tooling is available or the user says $reference-site-prototyping.

Lands finished Layline work in both repos it lives in, fullbuild.ai and the public mirror ryanportfolio/layline, with the drift and parallel-session checks the split copy needs. Use on $ship-layline or an ask to merge Layline into both.

Design, extend and audit fullbuild.ai visual surfaces. Use for new prototype identities under public/prototype or src/app/prototype, scroll narratives, GSAP/Lenis/R3F/SVG motion, visual-contract audits, or Playwright UI verification.

Publish a shipped prototype into the showcase at $prototype: rewrite, drafts row, and a finale card graded into the set. Use for $showcase-register, "add it to the showcase", "put it in the gallery", or a prototype missing from the index.

Design, extend and audit fullbuild.ai visual surfaces. Use for new prototype identities under public/prototype or src/app/prototype, scroll narratives, GSAP/Lenis/R3F/SVG motion, visual-contract audits, or Playwright UI verification.

Use proactively for hard layered work with dependent steps, load-bearing unknowns, repeated failures, or verification-sensitive handoff; also when the user asks for Fable mode.

Use when the user asks to review, audit, or stress-test recent code changes with fresh independent agents; requires exposed multi-agent tools or an authenticated Codex CLI.

Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy.

Use when the user asks to humanize, de-AI, de-slop, voice-match, or review prose for AI tells before publishing.

Use when the user asks to reduce per-turn context or token load, trim kernels, skills, or connectors, or propagate a generic context optimization to the starter.

Use when the user asks to humanize, de-AI, de-slop, voice-match, or review prose for AI tells before publishing.

Use when the user explicitly asks to lab or prototype a visual, UI, motion, or game-feel element with live tuning before production implementation.

Run the independent adversarial AI-review lane on a PR and post findings as a PR comment. Use when the user says $adversarial-review <PR#>, asks to run the review lane on a PR, or a PR still needs its independent review before merge.

Use for "automate me", "$automate-me", "create/update my -mode skill", or "turn my preferences into a skill". Mines the current project's transcripts plus direct questions, then drafts a personal <handle>-mode skill.

Verify a claim with fresh local evidence: restate it falsifiably, capture baseline and treatment, compare, return VERIFIED, NOT VERIFIED, or INCONCLUSIVE. Use for $verify-this, "prove it works", "did this fix it".

Use for "automate me", "/automate-me", "create/update my -mode skill", or "turn my preferences into a skill". Mines the current project's transcripts plus direct questions, then drafts a personal <handle>-mode skill.

Verify a claim with fresh local evidence: restate it falsifiably, capture baseline and treatment, compare, return VERIFIED, NOT VERIFIED, or INCONCLUSIVE. Use for /verify-this, "prove it works", "did this fix it".

Use when creating new skills, editing existing skills, or verifying skills work before deployment

Use when you have a spec or requirements for a multi-step task, before touching code

Mirror an existing external repo privately under the user's account and overlay the firmware: clone upstream, strip template-only files, privacy-sweep, run init-project. Use on $adopt-repo <url> or 'pull this repo into our firmware'.

First-principles chain: decompose, audit assumptions, recombine surviving blocks, test against reality; each step a fresh subagent fed only the prior artifact. Use on $dare, 'first principles', or 'are we solving the right problem'.

Coordinate parallel Claude Code sessions via a shared append-only Desktop HTML hub. Use on $session-hub, 'run parallel sessions on this', joining a hub, or proactively when edits or commits this session did not make appear in the checkout.

Mirror an existing external repo privately under the user's account and overlay the firmware: clone upstream, strip template-only files, privacy-sweep, run init-project. Use on /adopt-repo <url> or 'pull this repo into our firmware'.

First-principles chain: decompose, audit assumptions, recombine surviving blocks, test against reality; each step a fresh subagent fed only the prior artifact. Use on /dare, 'first principles', or 'are we solving the right problem'.

Coordinate parallel Claude Code sessions via a shared append-only Desktop HTML hub. Use on /session-hub, 'run parallel sessions on this', joining a hub, or proactively when edits or commits this session did not make appear in the checkout.

Use for "automate me", "$automate-me", or "turn my preferences / working style into a skill". Mines the current project's transcripts plus direct questions, then drafts a personal <handle>-mode skill.

Verify a claim with fresh local evidence: restate it falsifiably, capture baseline and treatment, compare, return VERIFIED, NOT VERIFIED, or INCONCLUSIVE. Use for $verify-this, "prove it works", or "did this fix it".

Use for "automate me", "/automate-me", or "turn my preferences / working style into a skill". Mines the current project's transcripts plus direct questions, then drafts a personal <handle>-mode skill.

Verify a claim with fresh local evidence: restate it falsifiably, capture baseline and treatment, compare, return VERIFIED, NOT VERIFIED, or INCONCLUSIVE. Use for /verify-this, "prove it works", or "did this fix it".