Find a listing

Managed shell-command runner for the machine Kiro Crew runs on: define command profiles (like your ninfer-serve line), start/stop them, watch live output, and cycle the local LLM backend without touching Kiro Crew.

Agent-neutral AI testing skills for Katalon True Platform covering the full software testing lifecycle: test planning, test design, test management and requirement traceability, test review, manual and automated execution with AI, failure analysis and release readiness, and test maintenance - operat

Drive a project end-to-end with BMAD-METHOD v6 (brainstorm to deploy) via a hybrid-autonomy orchestrator agent, with a multi-model council for reviews and second opinions.

The visual operating surface for AI-DLC Workflows: one Action Center for every approval gate, question and recovery decision across your repositories, with the evidence beside each decision. Studio reads AI-DLC's own files and sends every decision through the intent's own conversation; it never edit

View, edit, check and present HTML slide decks. Point at an element in the live preview, leave a comment, and let the agent apply the change.

Browse, edit, and comment on files on your gateway host — and let your agent do the edits you ask for. At its base it is a file browser and editor: walk any directory, open a file, edit it with syntax highlighting, autosave, and markdown preview. On top of that, select text and write a comment — lik

Kanban control tower for Dynatrace problems, services, and applications — with agentic smart ranking. Read-only, powered by dtctl.

Edits a todo.txt file: one task per line, no database. Point it at a folder you already sync, or use the one it creates. Priorities, projects, contexts and due dates colour as you type, and the format's semantics actually run rather than just highlight: rec: rollover on completion, t: thresholds, h:

Live an entire life inside an AI-simulated world that remembers what you did. Choose a path or write your own action; other people pursue their own goals, consequences can return years later, and every turn can reshape your relationships, status, and fate. Begin in the Age of Sword and Flame, explor

A live view of work across Kiro Crew that shows what needs you, what is running, and what recently finished.

Describe the deck you want in chat and get a real .pptx back. An agent interviews you, writes a brief, an outline and an art direction, then composes every slide and generates the file — and you watch each deliverable appear in a live preview beside the conversation. Slides render as animated SVG as

A personal advisor that researches real stores and helps solve problems that may or may not end in a purchase. The Preferences tab is your own reference list — paste what matters into chat. It never buys anything — you do.

A LaTeX paper editor with live PDF preview and an AI co-author. Write in a split-pane editor, press Cmd+S (Ctrl+S on Windows and Linux) to compile with pdflatex or tectonic, and read the rendered PDF beside your source. Compilation errors are parsed into a clickable list that jumps to the offending

Browse and read files on the machine running Kiro Crew without leaving the dashboard. Keep several files open in tabs, walk the directory tree with each file's git status shown beside it, search a whole folder for a string, and read markdown rendered rather than raw. It is read-only by design: it op

A design critique you can run before you ask a colleague. Give it one screenshot, a whole multi-screen flow, a Figma file, a git repo, or a URL that is already serving, and it finds the candidate screens, lets you choose what to look at, then reports back the way an experienced designer would: an ov

Local-first observability for your AI agent fleet: run history and health scores, day-by-hour heatmaps, P50/P95/P99 durations, failure diagnosis with probable-cause hints, work queue, scheduler view, live logs and real terminals. Observes the ecosystem it is pointed at (runs.db, queue/, schedule.jso

Shared markdown todo lists. Humans and agent sessions create ledgers, claim items, and complete them — one source of truth across all sessions.

A flag control tower that also does the code work: browse LaunchDarkly flags, read which ones are technical debt, and hand any flag to an agent to wire up or clean up.