Marketplace

Audit UIs against WCAG 2.2 AA and deliver prioritized, code-level fixes. Use when asked for an accessibility review, a11y compliance, or AT bug reports.

Analyze and reduce AI agent operating costs: model routing, prompt caching, context window management, and batch processing strategies.

Invoke when an autonomous task may outlast one context window—persist resumable checkpoints to a state file so work survives resets and interruptions.

Use when building or auditing an LLM-as-judge: calibrate it against human labels, measure agreement, and de-bias position and verbosity effects.

Invoke to audit or tighten an AI agent's tool/MCP permissions: diff granted access against observed usage and recommend a least-privilege allowlist.

Invoke when an AI system touches the EU market and you must classify its EU AI Act risk tier and output the obligations and article citations that follow.

Use before running untrusted or model-generated code, installs, or binaries: stand up a disposable, network-restricted sandbox, run it, then tear it down.

Invoke when building or tuning a RAG/semantic-search pipeline's chunking: pick strategy, chunk size, and overlap validated against retrieval metrics.

Use when a repo's test suite is red and you must drive it to green by patching the code under test within a budget, without masking regressions.

Use to scan agent transcripts, traces, and telemetry for leaked secrets or PII before they are stored or shared, and to fix the root cause.

Design feature flag strategies: flag types, rollout percentages, kill switches, technical debt management, and cleanup schedules.

Design secure execution environments for AI agents: Docker sandboxes, API key isolation, network policies, and resource limits.

Craft detailed agent personas with expertise areas, communication styles, decision frameworks, and behavioral boundaries.

Invoke when a downstream system must consume LLM output as JSON matching a fixed schema and malformed or missing fields would break the pipeline.

Use when a retrieval corpus has exact or near-duplicate docs/chunks inflating embedding cost or polluting top-k results, and you must detect and prune them.

Use to adversarially test an LLM app, agent, or tool pipeline for prompt-injection/jailbreak weakness and produce a scored hardening report.

Deploy and manage DigitalOcean infrastructure: Droplets, App Platform, Kubernetes, managed databases, spaces, and networking via doctl CLI.

Use when adding, upgrading, or auditing third-party packages: judge vulnerabilities, maintenance health, and supply-chain risk before they ship.

Design conversational analytics interfaces: NL-to-SQL, query disambiguation, result narration, and data storytelling.

Use when CI pipelines are flaky, slow, or failing intermittently; diagnoses root causes from logs and configs and prescribes durable fixes.

Optimize CI pipeline caching: layer caching, dependency caching, test splitting, and parallel execution strategies.

Design effective chatbot conversations: intent mapping, escalation flows, personality design, and fallback strategies.

Design multi-layer caching strategies: browser, CDN, application, and database caching with invalidation patterns.

Deploy and manage Azure infrastructure: VMs, AKS, App Service, databases, storage, networking, and Entra ID via az-cli with safety checks.