porter

Create, refine, and evaluate 3D assets, scenes, and renders for games, product visualization, concept development, animation preproduction, and marketing imagery. Use when work involves modeling, sculpting, retopology, UVs, materials, lighting, rendering, asset handoff, or visual polish of 3D content. Do not use for pure 2D brand design, generic illustration, or frontend/UI work unless the core deliverable is a 3D asset or render.

Plan and execute rigorous literature reviews, evidence syntheses, source evaluations, annotated bibliographies, research briefings, and academic-style analysis. Use when work requires scholarly source discovery, methodology-aware evidence review, comparison of studies, transparent inclusion criteria, citation discipline, or synthesis of peer-reviewed and institutional literature. Do not use for generic web browsing, marketing research, or opinion-first content that does not require academic rigor.

Audit, design, and improve digital products for accessibility and inclusive use across web, content, flows, and interface components. Use when work involves WCAG compliance, keyboard access, focus behavior, semantic structure, screen-reader usability, contrast, alt text, accessible forms, error handling, ARIA patterns, or remediation planning. Do not use for generic UI polish if accessibility is not the core concern.

Plan, assess, and improve technology-enabled agriculture systems spanning precision farming, crop monitoring, irrigation, sensing, decision support, farm operations, traceability, and agri-logistics. Use when work involves digital agriculture, climate-smart practices, sensor data, remote sensing, agronomic decision workflows, farm software, or deployment of technology in agricultural operations. Do not use for pure agronomy, policy, or generic logistics advice when technology systems are not central.

Review AI systems and outputs for safety, misuse, reliability, bias, privacy, security, and governance risks. Use when work involves model behavior review, prompt/agent safety, jailbreak resistance, hallucination risk, harmful-output analysis, evaluation design, red-teaming plans, or release-readiness assessments for AI features. Do not use for generic model training or analytics when risk review is not the core objective.

Design, build, and maintain trustworthy analytics models, metrics, semantic layers, and reporting-ready datasets. Use when work involves transforming messy source data into documented, testable, business-aligned models for dashboards, self-serve analytics, KPI definitions, or decision support. Do not use for raw infrastructure engineering or pure ad hoc analysis when the core problem is not analytics data modeling.

Lead animation work from concept through shot planning, animatics, blocking, review, polish, and final delivery. Use when work involves animation direction, motion performance, pacing, shot continuity, review notes, production sequencing, or animation pipeline decisions across 2D, 3D, product motion, or explainer content. Do not use for isolated illustration, static design, or 3D asset creation when motion direction is not the main problem.

Plan and write annual reports that combine strategy, performance, financial context, governance, impact, and forward-looking narrative into a clear stakeholder document. Use when work involves annual summaries for companies, nonprofits, startups, public-interest organizations, or internal business reviews. Do not use for generic blog writing, investor-relations legal advice, or standalone financial statement preparation.

Design HTTP and RPC APIs that are clear, versionable, secure, and easy for clients to adopt. Use when work involves endpoint design, resource modeling, request/response schemas, error contracts, pagination, idempotency, auth boundaries, webhooks, API migrations, or interface reviews. Do not use for pure backend implementation without contract work, frontend UI design, or infrastructure-only tasks.

Decide whether proposed changes require explicit approval before execution, especially for roster changes, role/permission changes, structural reconfiguration, destructive operations, policy exceptions, or other high-impact governance decisions. Use when work involves approving, denying, escalating, or framing guarded actions. Do not use for ordinary low-risk execution that is already clearly authorized.

Set visual direction for brands, campaigns, product surfaces, and asset systems so the work feels coherent, distinctive, and production-ready. Use when work needs concept-to-execution creative direction, style exploration, visual system alignment, critique, or final-call guidance across designers and asset makers. Do not use for single-asset execution without direction work, copywriting, motion production, or engineering implementation.

Define avatar identity direction for agents, characters, brands, and digital personas so profile art feels intentional, recognizable, and system-consistent. Use when work involves persona translation into visual traits, avatar style systems, iteration briefs, or critique of character/profile imagery. Do not use for generic logo work, broad brand strategy, or final illustration execution without direction responsibilities.

Design and implement server-side systems, APIs, jobs, data access, and service behavior that are correct, maintainable, observable, and production-safe. Use when work requires backend architecture within an existing system, endpoint or worker implementation, data-flow changes, debugging server behavior, or shipping tested backend code. Do not use for frontend-only work, pure API contract design without implementation, or infrastructure provisioning that does not materially change backend behavior.

Compare performance, cost, quality, maturity, or operational metrics against internal history, peers, competitors, or recognized standards to reveal where something is leading, lagging, or mismeasured. Use when work needs benchmarking design, metric normalization, peer-set comparison, target-setting, or interpretation of relative performance. Do not use for raw BI dashboarding without benchmark logic, experimental causality analysis, or generic market research with no comparative metric frame.

Turn business questions into trustworthy dashboards, reporting surfaces, KPI definitions, and self-serve analytics that operators and executives can actually use. Use when the main task is designing, critiquing, or standardizing BI outputs: dashboard structure, reporting semantics, drill paths, metric presentation, recurring performance reviews, or resolving conflicting reports. Do not use for broad analytical investigation, data-pipeline engineering, or predictive/causal modeling as the primary task.

Build rigorous, well-scoped bibliographies and annotated source maps for a research question, topic area, or evidence review. Use when the job is to find, assess, organize, and cite relevant literature or source material. Do not use for full literature synthesis, legal advice, or generic web search with no citation discipline.

Plan and write blog posts that are useful, readable, and strategically aligned to audience, channel, and business goal. Use when the task is long-form editorial content such as explainers, thought leadership, SEO articles, product education, or customer stories. Do not use for ad copy, social posts, or technical docs that need a different format.

Design cohesive brand identity systems, visual directions, and brand guidelines that create recognition and consistency across touchpoints. Use when the work is about logo systems, color and typography direction, visual language, brand refreshes, or codifying a brand for teams to use well. Do not use for one-off campaign art with no identity-system component.

Diagnose business problems, clarify stakeholder needs, and turn messy objectives into actionable requirements, process improvements, and decision-ready recommendations. Use when the work involves discovery, requirements definition, process analysis, KPI framing, stakeholder alignment, or prioritizing business changes. Do not use for pure financial modeling, pure product design, or implementation-only execution.

Forecast service and infrastructure capacity, model demand versus supply, and recommend headroom, scaling, and resilience plans before systems or teams hit failure. Use when the work is about growth planning, saturation risk, throughput limits, staffing or resource headcount tied to demand, or cost-versus-reliability tradeoffs in capacity decisions. Do not use for one-off incident response or detailed implementation of infrastructure changes.

Write evidence-backed customer stories and case studies that show challenge, solution, implementation, and measurable outcomes in a credible narrative. Use when the work is about customer success stories, B2B case studies, proof-driven marketing narratives, reference stories, or sales-enablement stories based on real results. Do not use for fictional storytelling, press releases, or generic blog posts.

Design and improve content delivery networks, caching rules, edge routing, origin shielding, asset delivery strategy, and CDN performance tuning. Use when the task involves cache hit ratio, latency reduction, edge behavior, invalidation strategy, multi-region delivery, WAF-at-edge tradeoffs, or diagnosing why static and dynamic content is slow, expensive, or incorrectly cached. Do not use for general backend performance tuning, database optimization, or application-level code refactors without a CDN focus.

Turn ambiguous chat requests into crisp execution moves, structured replies, routing decisions, and next-step orchestration. Use when the task is to keep a conversation moving efficiently, compress intent, decide what should happen next, coordinate parallel or staged work, or transform loose discussion into an action plan, delegation plan, or reply strategy. Do not use for deep domain execution where another specialist skill should own the work itself.

Improve chatbot behavior through prompt tuning, policy refinement, knowledge shaping, conversation review, rubric design, and training-data feedback loops. Use when the work is about making an assistant answer better, safer, more consistent, more on-brand, or more task-effective across real conversations. Do not use for model pretraining research, low-level ML infrastructure, or generic customer-support handling without a training or evaluation objective.

Design, audit, and improve CI/CD systems for build, test, release, deployment, rollback, and delivery governance. Use when the work involves GitHub Actions, GitLab CI, CircleCI, Jenkins, Buildkite, Argo CD, deployment workflows, release automation, trunk-based delivery, pipeline reliability, DORA metrics, supply-chain hardening, or faster safer shipping. Do not use for product feature implementation unless pipeline or release architecture is the main problem.

Design, evaluate, and synthesize clinical studies, protocols, and health evidence with attention to methodology, safety, endpoints, bias, and data integrity. Use when work involves clinical trial design, observational study design, protocol review, evidence appraisal, outcome interpretation, or health-data analysis in a clinical or translational context. Do not use for general wellness content, medical diagnosis, or patient-specific treatment advice.

Design, review, and improve cloud architectures for reliability, security, performance, cost, scalability, resilience, and operability across AWS, Azure, GCP, and hybrid environments. Use when the work is about service topology, environments, networking, data stores, identity, disaster recovery, multi-region design, cost tradeoffs, or platform patterns. Do not use for basic DevOps ticket execution or app-only feature work.

Implement clearly scoped code changes once the task domain is already obvious. Use when the job is execution-first: translating a defined requirement, ticket, bug report, or approved plan into working code within an existing codebase, and no domain-specific engineering skill is the better fit. Treat this as a fallback implementation skill, not the default for frontend, backend, or true fullstack work. Do not use for architecture discovery, broad redesign, code review, or specialist tasks that should route to a more specific skill.

Provide thorough, constructive review feedback on code changes with attention to correctness, maintainability, edge cases, tests, regressions, performance, and security. Use when the task is to evaluate a diff, pull request, patch, or implementation plan before merge. Do not use when the primary task is to write the code itself.

Delegate substantial coding work to an external coding harness or background implementation agent that can explore repositories, edit files, run commands, and iterate on code changes. Use when the task is implementation-heavy, benefits from autonomous repo exploration, or needs a dedicated coding lane rather than a normal conversational answer. Do not use for tiny inline edits, pure code review, or non-coding analysis.

Grow, moderate, and activate user communities across forums, groups, social channels, ambassador programs, and customer spaces. Use when the task is to shape community strategy, moderation policy, engagement programming, response playbooks, or health reporting. Do not use for one-off support troubleshooting, pure brand copywriting, or paid acquisition planning.

Analyze competitors, adjacent alternatives, and market moves to help teams position products, anticipate threats, and spot strategic opportunities. Use when the task is comparative market analysis, battlecard creation, pricing and packaging comparison, win-loss pattern synthesis, or strategic monitoring. Do not use for broad market sizing with no competitor lens, pure financial valuation, or internal product-roadmap ownership.

Gather, monitor, and synthesize competitive signals into ongoing intelligence that helps teams anticipate market moves, detect threats early, and support strategic, sales, product, and executive decisions. Use when the task requires structured monitoring, source tracking, signal triage, alerting, or intelligence operations across competitors and adjacent alternatives. Do not use for one-off competitor comparison when no ongoing intelligence system is needed.

Assess regulatory obligations, control gaps, monitoring needs, and remediation priorities across business operations. Use when work involves compliance programs, policy and control design, risk assessment, evidence collection, monitoring, audits, incident response, or preparation for legal/compliance review. Do not use for formal legal advice, litigation strategy, or attorney-only judgment.

Design, build, evaluate, and operationalize image and video analysis systems for detection, classification, segmentation, tracking, OCR, and visual understanding tasks. Use when the task involves computer vision pipelines, dataset curation, annotation strategy, model evaluation, deployment constraints, or real-world robustness issues. Do not use for generic ML tasks when visual data is not central.

Audit published or draft content for factual accuracy, structural weakness, messaging gaps, SEO issues, compliance risk, and conversion friction. Use when the task is to review existing copy, articles, landing pages, newsletters, product marketing, documentation, or campaigns and return a prioritized fix list or annotated rewrite plan rather than create net-new content from scratch.

Draft and rewrite general-purpose editorial, business, product, support, and internal written content when no more specialized writing skill is the better fit. Use when the main task is producing a clear, audience-aware first draft or clean rewrite for general communications, explainers, summaries, newsletters, product education, internal memos, or mixed-format content. Treat this as a broad editorial drafting skill, not the default for conversion copy, blog strategy, technical documentation, or UX microcopy.

Review contracts, terms, MSAs, SOWs, NDAs, DPAs, procurement documents, and clause drafts for risk, obligations, ambiguity, negotiation leverage, and operational impact. Use when the task is to spot issues in contract language, summarize key terms, propose redlines, or prepare negotiation comments. Do not use for formal legal advice, jurisdiction-specific legal opinions, or privileged attorney-only work.

Write persuasive, conversion-aware messaging for product, marketing, sales, lifecycle, and launch contexts. Use when the task is to sharpen value propositions, offers, headlines, landing-page copy, ad copy, email sequences, CTAs, campaign messaging, or objection-handling language intended to drive action. Do not use for broad editorial drafting, brand-identity design, or channel-strategy work when a more specialized skill fits better.

Reduce infrastructure, cloud, SaaS, vendor, and operational spend without breaking service quality or business goals. Use when the task is to analyze cost drivers, identify waste, prioritize savings actions, improve utilization, compare architectural tradeoffs, or connect spend to unit economics and accountability. Do not use for pure reliability engineering, procurement-law review, or accounting treatment work.