Claude Skills by curiositech

First-party planning doctrine for recording context clusters, hard artifact dependencies, order preferences, admission conditions, and a declared wave cap. NOT for HTP results or claims that file disjointness, parallelism, or human capacity are established theorems.

Use when designing or fixing a Playwright end-to-end test suite, debugging flaky tests, choosing locator strategies (getByRole vs CSS vs test-id), structuring fixtures and auth-state reuse, configuring parallelism and sharding, mocking third-party APIs via route(), or wiring trace-on-first-retry into CI. Triggers: "tests are flaky", page.waitForTimeout, page.locator with brittle CSS, login runs in every test, third-party API takes test offline, "should I use sleep here", parallel mode, shardi...

Plan and review Playwright screenshot workflows for visual regression, rendered-content checks, and human-auditable UI findings.

Create and scaffold plugin directories for Codex with a required `.codex-plugin/plugin.json`, optional plugin folders/files, valid manifest defaults, and personal-marketplace entries by default. Use when Codex needs to create a new personal plugin, add optional plugin structure, generate or update marketplace entries for plugin ordering and availability metadata, or update an existing local plugin during development with the CLI-driven cachebuster and reinstall flow.

Use when configuring a pnpm-workspaces monorepo, debugging hoisting/peer-dep resolution, adopting the catalog protocol, integrating with Turborepo or Nx, fixing "cannot find module" issues that only happen in CI, or migrating from npm/yarn workspaces. Triggers: pnpm-workspace.yaml setup, packageExtensions for broken peer deps, .npmrc tuning (public-hoist-pattern, shamefully-hoist), workspace protocol (workspace:*), filtering with --filter, version mismatches across packages, "Cannot find modu...

Applies political philosophy frameworks (Hobbes, Locke, Rousseau, Rawls, Sen, Ostrom) to software architecture decisions -- authority models, resource allocation, capability assignment, and commons governance. NOT FOR general philosophy courses, ethics of AI alignment, trolley problems, or moral philosophy of consciousness.

Voice and structure guide for Port Daddy expository writing — the kind of essay that takes a hard technical claim (formal verification, mechanism design, cryptographic protocol) and turns it into something an educated reader can follow with pleasure. Use when authoring docs/concepts/* or docs/tutorials/* pieces that explain how the Port Daddy whitepapers' formal proofs actually work. NOT for landing-page marketing (use port-daddy-marketing-copy), NOT for ADR bodies (those stay internal), NOT ...

Voice and copy guide for portdaddy.dev and Port Daddy marketing surfaces. Use when writing or editing any user-facing copy on website-v2: landing, docs, hero, CTAs, sidebar labels, page intros, ADR summaries, blog posts, email, social. NOT for internal docs (ADR bodies, code comments, CHANGELOG).

Library of 24 named, textured Port Daddy user personas (solo/indie developers, small teams, big-team engineers, enterprise decision-makers, hiring managers evaluating Erich Owens/Curiositech, friends of Erich, AI power-users, curious newcomers, and one adversarial skeptic) for use as test subjects in UX friction audits and product appeal analysis. Activate on "who would find this confusing", "test this against real users", "persona", "target audience", "who is this for", "would an enterprise ...

Choosing a Postgres connection pool mode that matches your workload — pgBouncer session/transaction/statement modes, the prepared-statement story (pgBouncer 1.21+), Supabase Supavisor's port-per-mode pattern (5432 vs 6543), AWS RDS Proxy pinning triggers, and `pg_stat_activity` diagnostics. Grounded in pgbouncer.org, supabase.com, AWS docs, and PostgreSQL docs. NOT for ORM-specific connection config, Postgres replication/read-replica design, non-Postgres poolers (PgPool for MySQL, ProxySQL), ...

Use when a Postgres query is slow, p99 latency regressed after deploy, EXPLAIN ANALYZE output needs interpretation, the planner is choosing the wrong index, dead tuples are blowing up, autovacuum is falling behind, pg_stat_statements review is needed, or query plans regressed between releases. Triggers: "Seq Scan on big_table", "Rows Removed by Filter: 3M", "could not extend file", parameter-sniffing surprises, missing indexes, prepared-statement plan-cache regressions. NOT for schema design ...

Designing Postgres Row-Level Security (RLS) policies for multi-tenant authorization, especially in Supabase / PostgREST stacks — `CREATE POLICY` syntax, USING vs WITH CHECK, PERMISSIVE/RESTRICTIVE merge semantics, the `(SELECT auth.uid())` performance pattern that turns 171ms scans into <1ms, indexes still required, role-based bypass via BYPASSRLS, security-definer escape hatches. Grounded in postgresql.org, Supabase docs, and Gary Austin's RLS-Performance benchmarks. NOT for: general authori...

Review task definition, constraints, authority, and feasibility before decomposition. Use typed hard blockers and decision-relevant clarification, with source-specific calibration for any scalar score. NOT for post-decomposition DAG checks, skill assignment, or latency-sensitive streaming.

Construct PowerPoint, Keynote, Reveal.js, or Marp decks that are both legible from the back of the room AND feel premium / crafted / expensive. Use for pitch decks, interview decks, board decks, conference talks, slidedocs, investor memos, lecture slides. NOT for logo / wordmark design (→ brand-guidelines), copywriting (→ copywriting), motion graphics for video (→ motion-design-web), live-stage stagecraft (out of scope), or live-presentation coaching (→ tech-presentation-interview).

Evaluate desirability, positioning, and emotional resonance for products AND for technical documents - landing pages, wireframes, papers, monographs, textbooks, and whitepapers. Assess whether someone will WANT this (not just be able to use it) via identity fit, problem urgency, trust signals, and - for anything whose price is measured in hours rather than seconds - return on effort: time to first insight, cost transparency, payoff visibility. Activate on "will they like it", "market position...

Reviews project plans, PRDs, prototypes, and vibe-coded app concepts for immediate product/user-needs gaps such as account creation, cold start, missing AI-provider fallback, support, pricing, permissions, and trust. Use when a plan needs a skeptical product reviewer before agents build it. NOT for code review, visual polish, or market research without a product artifact to inspect.

Facilitates bounded, consent-based discourse among participants with declared peer standing by selecting dialogue, debate, or deliberation; requiring reciprocal steelmans before critique; mapping claims, evidence, cruxes, and dissent; and closing in one explicit terminal state. Use for structured peer conversations. NOT for therapy, recovery counseling, abuse or coercive control, crisis support, legal or HR mediation, hierarchical disputes, argument-only analysis, discourse-suite orchestratio...

Audit consistency of structured declarations, not actual authorization or evidence truth. Prepare a scoped decision packet with evidence, authority, dissent and an impact preview. Use when a stale policy, cross-artifact consequence or multi-reviewer disagreement needs accountable synthesis. NOT for generic summarization, runtime administration, majority voting, authorization verification or automatically publishing a decision.

Expert prompt optimization for LLMs and AI systems. Use PROACTIVELY when building AI features, improving agent performance, or crafting system prompts. Masters prompt patterns and techniques.

Audits and designs evidence-bound pre-effect action-adjudication contracts: typed proposals, policy- and authority-bound decisions, one-use permits, effect receipts, and a claim ladder that separates verifier correctness from complete mediation. Use for a concrete consequential-action boundary, threat model, bypass inventory, or adjudication evidence packet. NOT for claiming a deployed reference monitor, generic policy-language authoring, generic Lean or Datalog work, prompt safety, post-hoc ...

Formal protocol verification with ProVerif and Tamarin Prover for multi-agent coordination systems. Use when modeling secrecy properties (escrow opacity, session note confidentiality), replay resistance (token dismissal finality, backup-restore attacks), delegation chain correctness, or append-only log integrity in applied pi calculus. Use when deciding between ProVerif and Tamarin for a given property. Use when extending existing .pv models (e.g., harbor_card_v3_delegation.pv) with new secur...

Progressive Web App development with Service Workers, offline support, and app-like behavior. Use for caching strategies, install prompts, push notifications, background sync. Activate on "PWA", "Service Worker", "offline", "install prompt", "beforeinstallprompt", "manifest.json", "workbox", "cache-first". NOT for native app development (use React Native), general web performance (use performance docs), or server-side rendering.

Advanced Python 3 patterns for modern coding interviews and production-quality refresh work. [What: dataclasses, descriptors, decorators, context managers, generators, structural pattern matching, walrus operator, typing, protocols, async basics, metaclasses, __slots__, attribute lookup, stdlib collections] [When: preparing for Python interviews after years away, translating Python 2-era or early Python 3 knowledge into current idioms, explaining metaprogramming, choosing Pythonic interview i...

Use when debugging asyncio event-loop hangs, blocking calls inside coroutines, asyncio.gather vs TaskGroup choice, cancellation handling, structured concurrency, mixing sync and async, run_in_executor decisions, asyncio task lifetimes, "Task was destroyed but it is pending" warnings, or async context manager bugs. Triggers: event loop blocked by sync I/O, ConnectionResetError on cancellation, asyncio in libraries that also offer sync API, FastAPI/aiohttp performance regressions, queue.Queue u...

Use when designing or fixing the retrieval side of a RAG system, choosing chunking strategy (fixed-size / recursive / semantic), implementing hybrid search (BM25 + dense) with RRF fusion, adding a cross-encoder reranker, evaluating with RAGAS, or running an index-freshness pipeline. Triggers: "RAG keeps citing the wrong doc", chunk size 512 tokens with overlap, RRF reciprocal rank fusion, dense+sparse hybrid, cross-encoder rerank top-5, RAGAS faithfulness / context precision, voyage-3-large v...

Formal framework for modeling intelligent agents through Beliefs, Desires/Goals, and Intentions as distinct mental states with rigorous semantics

Apply the Rao–Georgeff 1995 BDI interpreter and its representation choices

Interview techniques using rapport-building to improve information quality and yield from sources

Research linking rapport quality in interviews to information yield and intelligence value

Use when designing rate limiting for an API, choosing between token bucket / sliding window / leaky bucket / fixed window, implementing it in Redis, deciding edge (Cloudflare/Upstash) vs origin enforcement, sizing per-user vs per-IP vs per-endpoint quotas, returning the right 429 response with Retry-After, or fixing the boundary-burst bug in fixed-window limiters. Triggers: 429 too many requests, INCR + EXPIRE, ZADD + ZREMRANGEBYSCORE + ZCARD, X-RateLimit-Remaining header, Cloudflare WAF rate...

Drawing the `'use client'` boundary correctly in React Server Components apps (Next.js App Router, RSC frameworks) — leaf-pushing, slot composition, serialization rules, and environment poisoning prevention. Grounded in react.dev and Next.js 16 docs. NOT for: generic React patterns, data-fetching strategy inside Server Components, hydration-mismatch debugging, or client bundle performance profiling.

Builds DAG visualizations using ReactFlow v12 with custom nodes, ELKjs auto-layout, Zustand state management, and live state updates via WebSocket. Use when implementing workflow visualization dashboards, creating custom agent node components, integrating ELK layout algorithms, or wiring execution state into React components. Activate on "ReactFlow", "workflow visualization", "DAG visualization", "ELKjs", "custom nodes", "node-based editor", "graph visualization". NOT for writing Mermaid diag...

Generate legally-sound terms of service, privacy policies, and medical disclaimers for recovery and wellness applications. Expert in HIPAA, GDPR, CCPA compliance. Activate on 'terms of service', 'privacy policy', 'legal terms', 'medical disclaimer', 'HIPAA', 'user agreement'. NOT for contract negotiation (use attorney), app development (use domain skills), or moderation (use recovery-community-moderator).

Expert guidance for designing and implementing onboarding flows in recovery, wellness, and mental health applications. This skill should be used when building onboarding experiences, first-time user flows, feature discovery, or tutorial systems for apps serving vulnerable populations (addiction recovery, mental health, wellness). Activate on "onboarding", "first-time user", "tutorial", "feature tour", "welcome flow", "new user experience", "app introduction", "recovery app UX". NOT for genera...

Follow Recovery Coach codebase patterns and conventions. Use when writing new code, components, API routes, or database queries. Activates for general development, code organization, styling, and architectural decisions in this project.

Trauma-informed AI moderator for addiction recovery communities. Applies harm reduction principles, honors 12-step traditions, distinguishes healthy conflict from abuse, detects crisis posts. Activate on 'community moderation', 'moderate forum', 'review post', 'check content', 'crisis detection'. NOT for legal documents (use recovery-app-legal-terms), app development (use domain skills), or therapy (use jungian-psychologist).

Companion for finding the courage to quit drugs and rebuild a life after addiction stole your imagination, dreams, and executive function. Speaks to lived experience, not clinical distance. Covers rebuilding executive function, re-learning to dream, overcoming anhedonia, identity beyond 'addict', courage at each stage of change, and practical techniques for when everything feels impossible. Activate on 'courage to quit', 'can't imagine life sober', 'anhedonia recovery', 'can't feel anything s...

Write neuroscientific, peer-oriented drug education content that roots experiences in body/brain mechanisms. Use when creating educational articles, explaining neurological phenomena, demystifying recovery challenges, or answering "why does this happen?" questions. Activates for harm reduction content, psychoeducation, recovery science writing, and content that reduces shame through understanding.

Privacy-first social features for recovery apps - sponsors, groups, messaging, friend connections. Use for sponsor/sponsee systems, meeting-based groups, peer support, safe messaging. Activate on "sponsor", "sponsee", "recovery group", "accountability partner", "sober network", "meeting group", "peer support". NOT for general social media patterns (use standard social), dating features, or public profiles.

Facilitate multi-perspective synthesis of consequential documents through scoped research, independent contributions, evidence mapping, critique, consolidation, and owner disposition. Use when a document needs competing perspectives or an auditable account of unresolved choices. Activates on "synthesize document", "multi-agent authorship", "founding document", "architecture document", "constitutional document", or "multi-perspective document". NOT for ordinary single-author drafting, quick su...

Use when designing caching strategies (cache-aside, write-through, write-behind), implementing distributed locks, building rate limiters, leaderboards, real-time streams (XADD/consumer groups), pub/sub, or tuning eviction policies. Triggers: thundering-herd on cache miss, dogpile on key expiry, Redlock vs SET-NX-PX choice, sliding-window rate limiter, hot-key on a single cluster slot, big-key blowup, MULTI/EXEC across slots, KEYS in production. NOT for Redis Cluster operations/admin (differen...

Adversarial security review fleet for the Port Daddy whitepapers (Bonded Commons, Anchor Protocol). Use when a paper version is being prepared, when a coordination/cryptographic claim is being added, or when paired with the whitehat-defense skill in a versioned red-vs-white iteration round. NOT for production threat response — see SECURITY.md for incident handling.

Conducts thorough landscape research, competitive analysis, best practices evaluation, and evidence-based recommendations. Expert in market research and trend analysis.

Systematic research methodology for formulating problems, building evidence-based arguments, and communicating findings in academic contexts. For scholarly writing, thesis development, and knowledge contribution. NOT for creative writing, journalistic reporting, or business analytics.

Take a formal result to publication when it sits between fields — choosing a venue and positioning against the right literature, finding prior art that uses none of your vocabulary, making imported machinery graspable, and structuring the paper the way the target community expects. Use when drafting or revising a paper for EC/AAMAS/CAV/S&P/OSDI/POPL-class venues, writing a related-work or contributions section, checking whether a term or result is already taken, or preparing a submission-read...

The general retry/backoff/circuit-breaker canon is solved by [[circuit-breakers-and-retries]] — defer to it for jitter formulas, retry budgets, and the breaker state machine. THIS skill is only the load-once-dependency delta: wiring a breaker + backoff around a MEMOIZED lazily-loaded resource (native lib, ML model, connection pool, embedder) so a permanent failure is not cached as a rejected promise and re-awaited every tick. Use when a load-once dep can fail permanently, when you see a getX(...

Resolution-trace damping is an implementation-specific stigmergic heuristic for multi-agent systems: after an agent solves a problem at a node, it deposits a RESOLUTION trace there. An opt-in read helper can return a damped effective pheromone under the implementation's stated formula. This is a candidate for reducing repeated attraction to completed nodes, not a demonstrated prevention of redundant work or starvation. Verify source version and navigation behavior before claiming that agents ...

Design agents that make timely, good-enough decisions under time pressure, limited computation, and changing environments using commitments, partial plans, and filter-override control. Use when balancing deliberation against action, choosing commitment depth, or tuning reconsideration behavior in a bounded agent. NOT for classical full-information optimization, deterministic planning with fixed requirements, or cases where deliberation cost is negligible.

Designs and audits leveled, bounded, correlated logging for a long-lived service or daemon — a per-key rate-limit/dedup governor with suppression rollups, sampling that reports true totals, rotation plus the launchd/systemd captured-stdout trap, one-logger discipline, and observability that can never crash the process. Use when adding logging to a poll/retry/heartbeat loop, when logs eat disk or a daemon writes an unrotated file, when reviewing log level discipline, or when threading request/...

Reverse proxy architecture for AI agent systems — Nginx, Caddy, Traefik, and Cloudflare. Agent-specific patterns: routing to specialist agents, load balancing, TLS termination, rate limiting, WebSocket proxying, header injection for agent identity, and the "agent gateway" pattern. Covers LLM routers (LiteLLM, OpenRouter) as intelligent reverse proxies, service mesh (Envoy/Istio) for agent fleets, and MCP servers behind reverse proxies. Activate on 'reverse proxy', 'nginx proxy', 'caddy', 'tra...