7 Kiro Features You're Probably Not Using An AWS Developer Advocate highlights seven underused features of Kiro, an agentic harness for CLI, web, IDE, and iOS. The features include granular tool permissions via Tab key, a rich context menu in the IDE, and structured starting points like Bug Fix workflow that restricts file changes. Did you know that Kiro https://kiro.dev/?trk=1ad04439-1c50-4fdd-a845-d07d2655fe7a&sc channel=el doesn't just have a Spec-Driven Development SDD flow, but also a bug fix workflow that helps you resolve one issue at a time? That's one of seven features worth knowing about. If you're completely new to Kiro, it's an agentic harness for the CLI, web, IDE, iOS, and more. It helps teams and individuals do their best work while coding. I've been using it since it launched in July last year, and I keep finding features I didn't know were there. Full disclosure: I'm a Developer Advocate at AWS, and Kiro is a part of AWS. I use it every day, and I'll be forthcoming about the parts that are still preview or experimental. Heads up:Kiro ships fast. I've flagged the version-sensitive bits of these features inline. Check the docs if something looks different in your build. After talking to a lot of people about Kiro, one of the main pieces of feedback I get is on approving commands. When Kiro asks permission to run a shell command, the default reaction is to hit yes and move on. Then it asks again for the next git command. And the next one. Press Tab instead in the CLI. This allows you to edit it and put the exact permissions you'd like. For example you can be pickier on the trust tiers: git pull --rebase this exact command git pull git pull with any arguments git anything git the entire shell tool Whatever you pick persists for the session and gets stored as a regex in your agent's allowedCommands . There's also /tools trust-all , which trusts everything. It's the documented replacement for the old /acceptall , and the security docs https://kiro.dev/docs/cli/chat/security/?trk=1ad04439-1c50-4fdd-a845-d07d2655fe7a&sc channel=el are blunt about it: don't use it in production or with sensitive data, and you're responsible for whatever it does. One version note: on CLI v3 this moves to a permissions.yaml file, so the agent JSON advice above is v2. More on v3 in a minute. Full details: tool permissions https://kiro.dev/docs/cli/chat/permissions/?trk=1ad04439-1c50-4fdd-a845-d07d2655fe7a&sc channel=el menu is bigger than file in the IDE Type in the IDE chat and you get a list of context providers. file is the one I use a lot. The full list: codebase · file · folder · git diff · terminal · problems · url: · code: · repository · current · steering: · docs: · spec: · mcp: You can stack them in a single message. git diff problems review my changes is a different request than either one alone. The one I've been using more is the terminal . It hands Kiro your recent terminal output and command history, so you stop copying and pasting errors: terminal why did this build fail One thing I find handy is the " Currently Open files." However, keep in mind that Kiro pulls in your open files and their dependencies automatically without you asking. Full list: context providers https://kiro.dev/docs/chat/?trk=1ad04439-1c50-4fdd-a845-d07d2655fe7a&sc channel=el context-providers Also you can't drag a file in the IDE's own explorer to the chat window, however you can copy and past the location of the file and past that into the chat window as a workaround. Kiro gives you structured starting points, and it's tempting to just start typing into the chat box instead. Instead you can choose any of these in the IDE. Bug Fix is the one I'd try first. It writes down current behavior, expected behavior, and unchanged behavior. I like how it doesn't change a bunch of files it shouldn't. It basically tells the agent what it is not allowed to touch. On a feature spec you also get a choice between requirements-first and tech design-first. If you already know your architecture, pick design-first. There's no reason to answer requirements questions you already have answers to. Kiro activates a skill https://kiro.dev/docs/skills/?trk=1ad04439-1c50-4fdd-a845-d07d2655fe7a&sc channel=el automatically when your prompt matches the skill's description. That works, but I feel like it doesn't always automatically activate. You can invoke one directly instead. Skills in .kiro/skills/ or ~/.kiro/skills/ show up as slash commands, so a skill named pr-review becomes: /pr-review That loads the entire instruction file, not just the description. Works in the CLI and the IDE. Powers https://kiro.dev/docs/powers/?trk=1ad04439-1c50-4fdd-a845-d07d2655fe7a&sc channel=el are a different thing and worth knowing about separately. A power bundles MCP server config with steering and optional hooks, and Kiro loads it only when your task is relevant to it. Powers are keyword-activated, and they're free on every plan. Open Help About and check your version against kiro.dev/downloads https://kiro.dev/downloads/?trk=1ad04439-1c50-4fdd-a845-d07d2655fe7a&sc channel=el . Kiro's own changelog says auto-updates were paused for IDE 1.0.x and tells you to download the latest release directly from the site. So if your version doesn't match what's on the downloads page, that's why, and no amount of restarting will fix it. Reinstalling keeps your settings, extensions, and sign-in state. On the CLI: kiro-cli --version kiro-cli update kiro-cli doctor install and config problems And /changelog in the CLI inside a chat session shows the release notes inline, which is a nicer way to find out what shipped than scrolling a webpage. One subscription covers the IDE, the CLI, Kiro on the web https://kiro.dev/docs/web/?trk=1ad04439-1c50-4fdd-a845-d07d2655fe7a&sc channel=el , ACP-compatible editors, and automation in CI/CD. Same credit pool. I really like the web version. You kick off a session, it runs in a cloud sandbox, and it opens the pull request when it's done. Your laptop can be closed. There are also Automations, which are cron-scheduled agent runs that open PRs on their own. Web is still in preview, gated to Pro and above, and it needs a connected GitHub account. There's an iOS app too, but it's early access via TestFlight, Pro or higher, and it's a surface for starting and reviewing cloud sessions rather than an editor. No Android yet. Three of these are behind a flag or a toggle. CLI v3. Run kiro-cli --v3 and it runs alongside your existing 2.x setup without touching it. You get the spec agent in the terminal: kiro-cli --v3 /spec new my-feature Plus capability-based permissions.yaml , standalone hooks in .kiro/hooks/ .json , and markdown agent configs. See the v3 docs https://kiro.dev/docs/cli/v3/?trk=1ad04439-1c50-4fdd-a845-d07d2655fe7a&sc channel=el for the full list of features. Agent Focus in the IDE. Toggle it from the button in the top right. The layout flips: parallel sessions down the left, chat in the middle, specs and diffs in a panel on the right. It's a different job than editing files. You're directing several agents and reviewing what they produce. Settings, powers, MCP management, the terminal, and direct file editing all still live in the normal view, so you'll toggle back and forth. /goal. Give it an objective and a definition of done, and it loops implement-then-verify until the criteria are met: /goal Migrate the auth module to the new SDK. Done when all tests pass and there are no TypeScript errors. This is available in v3 CLI. Default is five iterations, --max