Restarts the Kiro Crew gateway from inside a session while keeping the conversation going.
kirodotdev/KiroCrew
Gives a coding agent a persistent workspace of skills so it can keep working, improve itself, and carry goals across sessions.
- 71 skills
- 4.7k stars
The 5 most installed
Ordered by installs on skills.sh- gateway-restart2 installs
- goal-loop2 installs
Starts a self-improving loop from a goal and folder, then runs it until the done criteria are met.
- grill2 installs
Questions you one branch at a time to align on a plan before committing to it.
- kirocrew-app-dev2 installs
Builds, packages and publishes Kiro Crew external apps, covering manifests, UI, crons and hooks.
- rubber-duck2 installs
Has a different-model listener adversarially interrogate your reasoning to catch gaps, inconsistencies and unsupported claims.
The other 66 skills
- self-nudge-loop2 installs
Keeps one long-lived session working toward a goal by scheduling its own next turns until a deadline.
- self-update2 installs
Checks for and applies Kiro Crew updates on request.
- browser-auth1 installs
Lets you sign in to a site yourself in the live browser view, then hands control back.
- feature-request1 installs
Collects a bug report or feature request and routes it to the right reporting channel.
- kirocrew-commands1 installs
Looks up Kiro Crew CLI commands and answers setup and getting-started questions.
- llm-council1 installs
Convenes subagents on different model families to synthesize, vote or debate a hard question.
- self-rca1 installs
Runs a blameless root-cause analysis of the agent's own mistake, with Five Whys and a durable lesson.
Scans your code files and lists testable surfaces with hypotheses for the agent to fix later.
Deploys your pre-built app to your own AWS account and returns a public HTTPS link.
Saves, versions and updates generated UI like widgets, HTML and markdown for reuse.
Classifies a failing test as a real bug, a flake, or an environment issue before fixing.
Keeps watching a pull request, CI run, ticket or deployment and reports when it reaches an outcome.
Explains why a safety policy blocked a tool call and takes the approved alternative path.
Records a browser flow as video and GIF to show animations and multi-step interactions.
Reads and controls native desktop apps by clicking, typing and scrolling through the accessibility layer.
Reviews your scheduled jobs and suggests cheaper modes to cut token spend, without editing them.
Turns the current session into a reusable skill, saved as a candidate unless you ask for live.
Builds or changes a Bot's Dashboard tab, wiring real data in through declared fields.
Scaffolds a dashboard page together with its data contract, provider and parity test.
Helps you read gateway debug tool answers and understand what 400, 403, 404 or 501 mean.
Publishes an HTML, markdown or widget artifact to a public HTTPS URL on your own AWS account.
Reviews a UI design against usability, accessibility and content heuristics with severity ratings.
Explains a topic, error or design decision tuned to one named audience's level.
Records a demo video of a web feature, either narrated with voiceover or as a silent clip.
Guides a deep-dive research session one focused question at a time, logging every step.
Runs a frontend design workflow: mockup options first, then build, evidence and usability review.
Breaks a large goal into ledger items, runs one session per item and verifies each result.
Points to the replacement skill for conducting work-ledger goals; this alias is deprecated.
Shows a live iOS Simulator in the dashboard while building or testing an iOS app.
Runs the full Kiro Crew pull request loop from commit through CI to auto-merge.
Develops and verifies the Kiro Crew source repo inside a git worktree, not the live gateway.
Analyzes missed review findings to close Code Review Sage's blind spots over time.
Guides motion, layout and accessibility when building animated or interactive widgets and dashboards.
Explains where a meeting's notes, diagrams, tasks and calendar data live and what states mean.
Designs and calibrates a trustworthy metric with guardrails before any optimization begins.
Generates the internal pet's upcoming behavior schedule of moves, moods and notifications.
Notifies you about due reminders in natural language and marks them done.
Partially adjusts the internal pet's behavior schedule when plans change.
Summarizes Slack unread mentions, channel digests, topics and threads.
Provides a catalog of feature tips to show new users during their first week.
Checks watch items against their last status, suggests actions and updates the watchlist.
Investigates alarms and incidents using an incident board and knowledge ledger, and records fixes.
Diagnoses and fixes LaTeX compilation failures so the PDF builds again.
Adds non-destructive margin and inline comments to a LaTeX document without rewriting it.
Proposes inline replacement text in a LaTeX document as accept-or-reject suggestions.
Polishes the English of a LaTeX passage as tracked suggestions without changing meaning.
Sets the conduct, quality principles and LaTeX house style for editing a .tex paper.
Diagnoses what you really need, then recommends products from stores you name, never buying.
Manages one issue or pull request pipeline on a repo, tracking whether each item is being worked on and solved.
Boots a Kiro Crew feature worktree as an isolated throwaway pod and proves it starts, authenticates and renders.
Creates or restyles a PowerPoint deck, building slides from text or a URL and designing a reusable slide style.
Drafts a release note section from a range of commits, grouped by what the reader gets.
Checks, launches, connects to and tears down a remote crew running on AWS Lambda MicroVM.
Runs a deep single-pass code review of a pull request, covering design reasoning and nine code-level dimensions.
Runs vulnerability discovery on one target as a supervised fleet of auditors and verifiers with human gates.
Explains what the session summary panel shows, what it costs and how to make summaries work well.
Drives the Kiro CLI spec workflow from requirements through design and tasks to execution.
Runs a goal as a team, deciding what to keep, dispatching the rest and keeping the status board honest.
Applies batched visual edit requests made through the Design Tweak app to the page.
Opens a real external web page that the user watches live and can take over with one click.
Opens a preview card in the right panel for a local server URL you started.
Verifies your own UI change on an isolated local server by capturing and showing real screenshots.
Renders rich HTML inline that inherits the dashboard theme instead of clashing with it.
Writes side-effect-free backend pytest tests for the Kiro Crew repo that do not flake.