These docs ship inside the Python package. They are the end-user documentation:
in-app reading, dashboard Settings links, and the feature-tips catalog all resolve
here. Contributor and architecture docs live in ../../../docs/.
index.md is the in-app entry point. This README is the same set, organized for someone browsing the repository.
| Doc | Covers |
|---|---|
| getting-started.md | Install, first run, and background operation. |
| configuration.md | Config file reference, environment variables, and sandbox modes. |
| use-cases.md | Real-world workflows. |
| troubleshooting.md | Common problems and fixes. |
| blocked-commands.md | Why a command was refused, what the agent is told to do instead, and how to check your credential setup. |
| Doc | Covers |
|---|---|
| agents.md | Switching between specialized agents per conversation, thread, or cron job. |
| crew-members.md | Named crewmates, their standing DM threads, and routing work to one with select_crew / route_crew. |
| remote-crew.md | Reaching other machines running their own Kiro Crew gateway: the transports, the pane switcher, sessions that run on one, and federated history search. Formerly remote-instances.md, which now points here. |
| agent-spec-fields.md | Every agent-spec field, what it does, and how that differs per ACP backend. |
| skills.md | Drop-in markdown knowledge packs for domain-specific workflows. |
| steering-and-hooks.md | Steering files and their inclusion modes, the prompt library, and chat lifecycle hooks. |
| monitoring.md | Token-efficient pull-request monitoring and finite legacy fallbacks. |
| cron-and-scheduling.md | Scheduling recurring tasks. |
| subagents.md | Spawning parallel background workers for fan-out work. |
| dynamic-subagent-sizing.md | How the concurrent sub-agent cap is sized from host memory and CPU. |
| task-runner.md | Autonomous multi-step execution from a spec file. |
| research-lab.md | Multi-cycle research campaigns with exportable reports. |
| memory-and-learning.md | Persistent preferences, project context, and learned corrections. |
| knowledge-library-how-it-works.md | How the knowledge graph is built from your documents. |
| dashboard.md | The web dashboard: multi-session chat, memory management, live metrics. |
| issue-radar-pipeline.md | Which step of automated triage every issue is sitting in, how long it has been there, and what each agent session cost. |
| agent-questions.md | Letting an agent pause mid-turn to ask a clickable question. |
| followup-suggestions.md | Agent-proposed next steps above the composer. |
| feature-tips.md | Personalized tips pointing at features you have not used. |
| feature-videos.md | Short intro clips for features this install has not used yet. |
| inbound-webhooks.md | Letting external systems trigger an agent turn over HTTP. |
| deploy-web.md | Publishing artifacts to a public HTTPS URL on your own AWS. |
| snapshot-and-restore.md | Backing up and restoring Kiro Crew state. |
| workflows.md | Multi-phase agent runs you can watch, restart in part, and save for reuse. |
| secrets-vault.md | Storing credentials encrypted where the agent cannot read them. |
| monitor-loops.md | Keeping one session checking something on an interval until an exit condition fires. |
| session-ledger.md | The durable per-session work record that survives context compaction. |
| session-control.md | Opening, forking, seeding, watching, stopping and closing another session, and the sidebar folders and tags that file them. |
| work-ledger.md | The conductor/worker record: work items, acceptance conditions, and how a dispatched worker reports. |
| artifacts.md | Saving, versioning, and reverting generated UI and documents. |
| computer-use.md | Reading and driving native desktop applications; opt-in and off by default. |
| decisions.md | Jev decisions: letting a small fast model pick the automatic skill for a sampled conversation and pick whether a mid-turn message steers or queues, with the shipped behaviour as the fallback; flagging a risky tool call on its own card in a session that approves its own calls, which changes no permission; and a basic diagnostic log for all of them. |
| browser-control.md | Driving a real web page from the dashboard's Browser panel. |
| Doc | Covers |
|---|---|
| slack-integration.md | Slack DMs, tool approval, streaming, channel monitoring. |
| discord-integration.md | Discord setup and behavior. |
| telegram-integration.md | Telegram setup and behavior. |
| teams-integration.md | Microsoft Teams setup, including Azure Bot registration. |
| webex-integration.md | Webex setup and behavior. |
| wecom-integration.md | WeCom setup and behavior. |
| weixin-integration.md | Weixin setup, and the risks to read first. |
| whatsapp-integration.md | WhatsApp (QR-linked personal account) setup, and the risks to read first. |
| feishu-integration.md | Feishu (Lark/飞书) setup and behavior. |
| imessage-integration.md | iMessage setup and behavior on a Mac that owns the Messages database. |
| channel-capabilities.md | One matrix of what every channel can do: streaming, buttons, uploads, reply length, approval timeout. |
| Doc | Covers |
|---|---|
| apps.md | The apps that ship in the package, the App Store that enables and updates them, and what enabling one grants. |
| mcp-apps.md | Rendering interactive MCP tool output in chat: the two gates, what a server declares, and the plain-text fallback. |
| connections.md | MCP servers and OAuth'd services: the curated catalogue, adding your own, credential custody, health badges, tool renames, and quarantine. |
| settings-deeplink.md | Answering "where is that setting?" with a link that opens and flashes the control, and the generated registry it comes from. |
Three constraints make this tree different from docs/:
- Filenames are an API.
tips.pyglobs*.mdhere and filters throughtips_allowlist.py, extracting each doc's H1 and first paragraph into the in-app feature catalog; dashboard Settings panels hardcode GitHub URLs to specific filenames; and a test pins one name. Renaming or deleting a file here is a code change, and a doc dropped from the catalog fails silently.scripts/docs-lint.shpins the coupled names. - The tree is flat, deliberately.
setup.cfg'spackage_dataglob for this directory does not recurse, so a file in a subdirectory would ship in the sdist but be missing from the wheel. - Not every file here is prose.
settings-registry.generated.jsonis a build artifact of the dashboard (npm run gen:settings), shipped beside settings-deeplink.md so the agent can enumerate the Settings controls at runtime. Do not hand-edit it; a frontend test byte-matches it against the live panels.
Because every doc here reaches every user, keep the content task-oriented and free
of internal design narration. An engineering note belongs in
../../../docs/ instead — for example
how the model's context is assembled,
which cites private symbols and so lives there rather than here. Each doc's first paragraph is
read verbatim as a feature description, so write it to stand alone.