Shape agent capabilities into your workflow.

principal-architect-review reverse-engineers a C4 model, identifies significant architecture decisions, and requires genuine alternatives, trade-offs, consequences, and reversibility in each ADR. A skeptical review probes coupling, operational cost, one-way doors, and simpler options, then turns agreed layering, cycle, import, and complexity rules into self-tested CI checks that expose drift.

ai-rmf-governor applies the NIST AI RMF Govern, Map, Measure, and Manage functions to an LLM or ML feature. It assigns ownership, records intended use and foreseeable misuse, names failure modes, sets metrics and acceptable thresholds, and scaffolds an offline evaluation gate. Model, prompt, use-case, or incident changes reopen the relevant work and require explicit residual-risk sign-off.

production-readiness-gate runs a phased Production Readiness Review before launch. It scopes blast radius, selects a profile, checks reliability, security, operability, cost, and observability, and runs offline probes for secrets, exposed endpoints, health checks, irreversible migrations, rollback gaps, and unbounded cost paths. A deterministic verdict yields GO, NO-GO, or GO-WITH-RISKS, with dated PRR records and tracked risk acceptances.

risk-register-csf builds a compact security program from an asset inventory, a NIST CSF 2.0 self-assessment, a living risk register, and a lightweight incident-response plan. It scores likelihood and impact deterministically, assigns owners, treatments, target states, and review dates, and keeps required runbooks for data exposure and account compromise. Re-assessment diffs prior artifacts and forces overdue open risks back to an explicit decision.

trading-ops-manager is built for Indian NSE/BSE swing traders who want one conversational place to log trades, track paper ideas, manage X drafts, close positions, and review performance. It stores the journal in Supabase and uses INDmoney for live prices and calculations such as risk, returns, and behavior patterns. Supabase and INDmoney MCP connections are required; the trader still places broker orders.

crypto-legal-skill gives Solana founders structured reference material for incorporation, token launches, KYC/AML, stablecoins, DePIN, and crypto taxes across several jurisdictions. It adds phased checklists for token launches, VASP/CASP registration, and India-specific compliance, plus a legal-advisor workflow. The project is general information rather than legal advice, so current rules and binding decisions still require qualified counsel and official sources.

print-onepager-skill builds print-ready A4 one-pagers, sell sheets, and infographics with an institutional sand visual system. Its starter template fixes the hero and footer while the middle flex area absorbs remaining space, then measures the rendered page to catch overflow before PDF export. It includes reusable HTML, design tokens, and print settings for teams that need a clean single-page handout.

ux-writing-kr creates and critiques Korean UI microcopy for buttons, errors, empty states, onboarding, toasts, paywalls, dialogs, placeholders, and tooltips. It can return several variants with a recommendation and explain issues such as translationese, inconsistent honorifics, awkward length, or ambiguity. The skill emphasizes natural Korean service language and reassuring copy for moments that make users uncertain.

self-evolve provides a methodology and deterministic harness for improving a skill, repository, or project over repeated cycles. The loop reflects, proposes, evaluates, judges, and accepts changes, while code—not the model—decides acceptance, rejection, rollback, and which signal source to trust. It supports program checks, anchor verification, and generative evaluation for open-ended work where ground truth is limited.

Pantheon convenes six isolated reviewers for architecture, security, product fit, implementation truth, accidental misuse, and research claims. They review independently, surface clashes, resolve disagreements in proportion to the project’s stakes, and return a final verdict with a mandatory dissent log. Individual security, QA, or source-truth lenses can also be run without the full council.

Soundings structures personal decisions that cannot be solved by looking up more facts, such as career direction, founder fit, or family priorities. You state the question, select two or three advisors, receive independent views and one question from each, then get a synthesis of conflicts and the underlying values fork. The session ends with one user-chosen next step, not a recommendation.

home-auto-insurance-quotes connects an agent to a network of licensed independent agencies for live home, auto, and bundled insurance quotes. It supports conversational intake or an agent-driven quote-to-bind handoff, while the agency’s licensed human finalizes the policy. Coverage is focused on Texas today and eligibility is checked by state and product; paid rating and personal-data tools require a partner API key.

honestweek turns a completed week of local AI coding sessions into a shareable work report that includes work not yet visible in commits. It re-checks every git-verifiable claim against commits or aborts, labels each item as shipped, in progress, or designed but unproven, and attaches a source pointer. The draft remains private for the user to review and publish; there is no telemetry or automatic publishing.

legal-launch-review checks Chile-focused AI, SaaS, RAG, API, and web projects for personal-data handling, privacy notices, terms, basic security, AI risks, e-commerce duties, and licenses. It reviews the actual code and returns a simple red-yellow-green launch view with located findings and suggested fixes. The scope is advisory and limited to Chile, so laws, citations, and launch decisions must be verified with official sources or counsel.

sync-mac-windows keeps the portable parts of ~/.claude aligned between Windows and macOS using a Git repository plus a OneDrive mirror. It translates paths, preserves OS-specific MCP forms, and stages only sanitized skills, agents, commands, memory notes, and shared settings while excluding tokens, sessions, caches, and machine-only files. Audit, stage, apply, and mirror scripts default to dry runs.

universal-audit-skill classifies almost any work product, selects relevant verification domains, and produces a numbered audit record with severity-rated findings and remediation links. It supports code, applications, APIs, architecture, infrastructure, reports, spreadsheets, and media across Claude, Codex, Cursor, Gemini, and Copilot. Constructive and adversarial passes can be combined, while external standards and facts still need current validation for the project.

mihomo-skill is a Claude Code guide for configuring Mihomo, also known as Clash Meta, from basic rules to advanced features. It includes four scenario templates, a three-level DNS architecture guide, and troubleshooting coverage for common failure classes. Invoke it with /mihomo when you want a focused configuration reference rather than a separate proxy application.

loupe turns a Claude Code PR review into an interactive browser session. You keep the findings that matter, add your own notes, and choose the verdict; the tool can place Mermaid architecture or journey diagrams above the diff and link UML elements to changed lines. It accepts a GitHub PR or any local unified diff, with Python's standard library as its only required runtime.