Marketplace
Skills
Skills for Haarness: instructions and scripts an agent picks up when a task calls for them, added with one command.
Firebase AI Logic
Wires Gemini model calls into Android, iOS, web or Flutter apps through Firebase AI Logic, with App Check protection.
AI Search Optimization
Shapes content so AI answer engines can find, quote and cite it, and tracks how often a brand appears in AI answers.
Marketing Video Production
Plans and produces marketing videos with AI avatar tools or code-based frameworks such as Remotion.
A/B Testing
Designs experiments with a hypothesis, sample size and success metric, and builds an ongoing testing program.
Azure Preflight Validation
Runs step-by-step preflight checks on azd, Bicep or Terraform projects, including RBAC and identity, before an Azure deploy.
Azure Production Diagnostics
Troubleshoots Azure App Service, Container Apps, Functions, AKS, VMs and messaging with logs, metrics and resource health.
Git Safety Net
Audits branches, stashes, worktrees and clones for unpushed work, recovers lost commits and proves branches merged before cleanup.
Tech Selection Checklist
Runs a technology choice through framing, prior-art search, evidence probes and pass/fail filters, then states who decides.
Prisma Postgres Setup
Gives an app a Postgres database on Prisma's hosted service, reusing an existing one when present, and stores the URL safely.
Query PostHog Data
Teaches the agent when to use PostHog's typed trend, funnel and retention queries versus HogQL, with schemas and worked examples.
Investigate Metric Changes
Root-causes a dropped or spiking product metric with breakdowns, prior-period checks and flag or deploy correlation.
QA Process Kit
Bootstraps a QA folder with test-case templates, tracking sheets and reports, and guides test cases, bug severity and OWASP checks.
ARM and Bicep to Pulumi
Rebuilds Azure infrastructure defined in ARM or Bicep as a Pulumi program and adopts the live resources without recreating them.
CloudFormation to Pulumi
Ports AWS CloudFormation to Pulumi resource by resource, keeping logical IDs as names so live resources can be adopted.
Terraform to Pulumi
Moves a Terraform or OpenTofu setup to Pulumi, carrying over state so nothing is rebuilt, then rewrites the code to match.
Migrate Discovered Stacks
Turns a cloud stack found by Pulumi Cloud Discovery into Pulumi code by triaging, bulk-importing and reconciling its resources.
AWS CDK to Pulumi
Ports an AWS CDK app to Pulumi by synthesizing its templates, converting each resource and optionally importing state.
Design Token System
Builds primitive, semantic and component token layers as CSS variables, checks code for hard-coded values, and drafts token-based slides.
UI/UX Pro Max
Searches a bundled database of styles, palettes, font pairs, UX rules, charts and stack notes to set a design direction and check UI work.
Brand Asset Design
Produces logos, identity mockups, SVG icons, banners and social images, with artwork generated through third-party image models.
Brand Guidelines
Defines brand voice, messaging and visual identity in a guidelines file, then checks assets and syncs colours into design tokens.
Avoid AI Writing
Finds and fixes machine-sounding prose in docs, READMEs and posts, in detect, rewrite or edit-in-place modes.
Dev Browser
Drives a real Chrome through short Puppeteer scripts with named pages that stay open between commands, for testing and web tasks.
Agent Tool Design
Writes tool descriptions, schemas, response formats and error messages that agents can choose and recover from reliably.
Multi-Agent Patterns
Helps decide whether several agents are worth it and how to structure supervisor, swarm or handoff designs with isolated context.
LLM Project Planning
Checks whether an LLM fits the task, then shapes the pipeline and estimates token use and cost.
Long-Horizon Task Briefs
Writes launch prompts for long-running or parallel agents with defined terms, an exact success test and handoff contracts.
Agent Memory Systems
Designs persistent agent memory with entity tracking, time-bounded facts, graph or vector lookup and consolidation.
Context Optimization
Cuts token cost through context budgets, observation masking, cache-friendly prompt prefixes and partitioning across sub-agents.
LLM-as-Judge Evaluation
Builds LLM-judge scoring with pairwise comparison, calibrated rubrics and checks for position and length bias.
Hosted Agent Infrastructure
Plans background coding-agent infrastructure: sandboxes, warm pools, snapshots, session persistence and multiplayer sessions.
File-Backed Agent Context
Moves agent state into files: scratchpads, offloaded tool output, handoff notes and cleanup rules for context kept outside the prompt.
Agent Harness Engineering
Designs the control loop around autonomous agents: fixed evaluators, narrow edit areas, persistent logs, rollback and approval gates.
Agent Evaluation Systems
Sets up rule-based checks, regression sets, rubrics and release gates to measure agent pipelines over time.
Questionnaire Builder
Asks who you are sending to and what you need back, then writes a focused markdown questionnaire for that person.
Context Compression
Designs compaction and structured summaries for long agent sessions so key choices, touched files and open risks survive.
Context Degradation Diagnosis
Diagnoses lost-in-the-middle, poisoned, clashing or confusing context in agent runs and suggests ways to contain each failure.
Writing For Agents
Explains how to word and structure skills and agent instruction files so the right material loads and behaviour stays consistent.
Plan To Tickets
Splits a plan into vertical-slice tickets with explicit blockers, reviews the breakdown with you, then files them in order.
Issue Triage
Keep an issue tracker under control by sorting new issues and PRs and writing briefs an agent can pick up.
Setup Wizard Generator
Generates an interactive bash script that walks a person through manual setup steps and saves the values they paste in.
Wayfinder Planning Map
Break a project too large for one sitting into a tracked set of decisions and work through them over time.
Conversation To Spec
Condenses an agreed discussion into a structured spec with user stories and test seams, then files it in your issue tracker.
Teaching Workspace
Runs a multi-session course in the current folder with a mission, short HTML lessons, quizzes, references and learning records.
Session Retrospective
Looks back over a coding session and proposes changes to checks, standards, steering files and tooling so the next run goes better.
Pre-Commit Hooks Setup
Adds Husky and lint-staged to a JavaScript repo so commits are formatted, type-checked and tested automatically.
Test-Driven Development
Guides red-green cycles one vertical slice at a time, with tests agreed at public seams and common test anti-patterns called out.
Throwaway Prototype
Builds disposable code to settle one design question, either a clickable state-machine demo or several switchable UI variants.