* fix(auth): scope session-keyed reads and mutations to the workspace owner Session-keyed service entry points did not apply per-user scoping, so knowing another user's session or task ID was enough to reach it: - GetTaskSession, ListTaskSessions, GetPrimarySession returned rows for any ID. These feed GET /task-sessions/:id and GET /tasks/:id/sessions, whose DTO carries session Metadata, the host worktree path, the branch, and container/executor IDs. - DismissLastAgentError mutated any session. - ApproveSession approved a workflow review step on any session, which advances the task's workflow — the highest-impact of the set. Authorize via authorizeTaskID off the row already read, rather than AuthorizeSessionAccess, which would re-read the session on what is a hot path. Internal identity-less callers (pollers, event bus, office schedulers, office agent JWTs) and the synthetic identity used while auth is disabled stay unscoped, and unowned pre-auth rows stay visible. Batch helpers that take ID lists are deliberately left alone: their callers derive IDs from an already-authorized task list, so per-ID checks would add N queries to list views for no gain. Recorded in AGENTS.md. Also route httpApproveSession through handleNotFound so a denial answers 404 like every other session route instead of a 500 that would confirm the session exists. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(auth): scope the session-keyed paths that bypass the task service An adversarial sweep of the previous commit found session-keyed surfaces reachable with a caller-supplied ID and no ownership check. They bypass the task service, so scoping its methods did not cover them: - POST /task-sessions/:id/{set-mode,set-model,set-config-option, authenticate} and .../processes/:processId/stop resolve their execution with a bare in-memory lookup, which skips the GetOrEnsure* chokepoint where the lifecycle per-user check runs. A non-owner could switch another user's model or mode, re-authenticate their agent, or kill their process. Guarded by a shared denySessionAccess helper. - The orchestrator reads sessions through its own repo handle, so GetTaskSessionStatus and CheckSessionPR inherited nothing from the task service. CheckSessionPR also installs a PR watch as a side effect. Added a SetSessionAccessChecker hook mirroring the lifecycle manager's, wired to taskSvc.AuthorizeSessionAccess. The check runs before CheckSessionPR's unconfigured-GitHub early return so it applies regardless of setup, and denials reuse the "session not found" shape so they cannot be told apart from a missing session. - GetMessage had no scoping while ListMessages beside it did, so a caller holding another user's (session_id, message_id) pair could read their shell output. This does NOT close the whole surface. A second sweep found the orchestrator's session-mutating WS actions (session.stop/delete/rename/ set_primary/set_plan_mode/reset_context/recover/launch/ensure, agent.cancel, permission.respond) and agent.stdin / agent.resize also unscoped — agent.stdin writes arbitrary input into another user's agent process. Those touch live session lifecycle and need their own change; recorded in AGENTS.md so the gap is not mistaken for covered. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(auth): scope every session-keyed WS action, plus a gateway backstop Completes the sweep. Dispatched WS actions were left to authorize themselves at the service layer, and a dozen never did — a caller with a guessed session ID could stop, delete, rename, relaunch or reset another user's live agent session, cancel their in-flight turn, answer their permission prompts, or write arbitrary bytes into their agent process. Two layers, deliberately both: 1. Gateway backstop. Client.authorizeAction checks any dispatched payload carrying task_id or session_id against the client's identity before dispatch. Per-method opt-in across ~74 actions in four packages had been missed repeatedly; this makes a newly added action scoped by default. Verified safe to apply generically: session_id means a task session everywhere, and user-shell/terminal actions key off task_id / task_environment_id, so nothing is misread. 2. Explicit guards, for defense in depth and for non-WS callers: - orchestrator: StopSession, DeleteSession, SetPrimarySession, RenameSession, ResetAgentContext, CancelAgent, RespondToPermission, SetSessionPlanModeByID, RecoverSession, LaunchSession, EnsureSession, CancelTransientRetry. Added SetTaskAccessChecker for the task-keyed ones. LaunchSession guards every intent at the single funnel; SessionID is empty on create, so the task check carries that case. - agent.stdin / agent.resize now call lifecycleMgr.CheckSessionAccess, the helper already documented for bare-lookup handlers. Guards go first in each method, before any repo/executor use. TestSessionKeyedEntryPointsGuardBeforeDependencies runs all twelve with nil dependencies and fails on a panic if one is placed too late. Denials are indistinguishable from a missing resource: WS returns not-found, CancelTransientRetry reports "nothing cancelled", and GetTaskSessionStatus reuses its "session not found" shape. Unwired checkers stay permissive throughout, so identity-less internal callers (schedulers, event bus, workflow engine) and pre-auth single-user instances are unaffected — pinned per entry point. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(auth): scope environment-keyed user-shell actions Third sweep finding, in surface the previous commit's own comment claimed was covered. The user-shell actions key off task_environment_id and treat task_id as optional, so a task_id-only backstop missed them: user_shell.stop with an empty task_id and a foreign task_environment_id fell through to the legacy runner path and tore down another user's terminal PTY. That path never reaches GetOrEnsureExecutionForEnvironment, where the environment check normally runs, so neither layer caught it. - Gateway backstop now also reads task_environment_id, gated by a new AuthPolicy.ActionEnvironment hook wired to taskSvc.AuthorizeEnvironmentAccess. This covers all three environment-keyed actions generically, not just the reported one. - wsUserShellStop guards explicitly via a new lifecycleMgr.CheckEnvironmentAccess, mirroring CheckSessionAccess, so the fallback is safe independent of the gateway. Corrected the comment that wrongly asserted task_environment_id was already covered, and recorded in AGENTS.md both the exact field names the backstop reads and why a permissive type on one of them would silently opt out. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(auth): authorize both IDs on task+session entry points Review finding (P1). Entry points taking both a task and a session ID only checked the session, so a caller could satisfy that check with one of their own sessions while pointing taskID at another user's task: - CheckSessionPR then used the unchecked taskID for GetTaskPR, resolveTaskRepo, resolvePrimaryTaskRepositoryID and EnsurePRWatch — cross-workspace PR disclosure plus a PR-watch write. - CancelTransientRetry handed it to handleRecoverableFailure, which writes against that task. Both now use a new authorizeTaskSessionPair, which authorizes each ID and additionally requires the pair to be consistent, so the two arguments cannot describe different tasks. GetTaskSessionStatus already validated the pair after loading the session and is unchanged. The consistency half is skipped when the session cannot be loaded: both IDs are authorized by then, so a mismatch is a caller bug rather than a cross-user leak, and failing there would change behavior for identity-less internal callers. Recorded the rule in AGENTS.md — a two-ID entry point must never use authorizeSession alone. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(auth): scope AttachWorkspaceSources Found while rebasing onto main. POST /tasks/:id/workspace-sources and the add_workspace_sources_kandev MCP tool landed on main during this PR's review with no ownership check, so a caller could attach repositories and folders to another user's task and have them materialized into that task's workspace. Same class as the rest of this PR, so fixed here rather than left for a follow-up. The WS dispatch backstop added earlier already covered the browser path (the payload carries task_id); this closes the HTTP route and the MCP tool, which do not go through it. ErrTaskNotFound already maps to 404 in workspaceSourceHTTPStatus, so the denial surfaces correctly without touching the handler. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(auth): map denied shell-output reads to 404 Review finding. GetMessage became per-user scoped earlier in this PR, so a denial now arrives as ErrTaskNotFound. httpGetShellOutput only mapped sql.ErrNoRows to 404, sending everything else to the 500 branch — so a foreign message answered 500 while a missing one answered 404, which is an existence signal and inconsistent with the 404-everywhere rule the other session routes follow. Kept the sql.ErrNoRows branch explicit rather than replacing it outright: isNotFound classifies typed sentinels and the "not found" substring, and "sql: no rows in result set" matches neither, so routing that case through handleNotFound alone regressed the missing-message response to 500 (caught by TestHTTPShellOutputSnapshot). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * test(auth): make the shell-output scoping test actually exercise the guard Both reviewers caught that TestHTTPGetShellOutputDeniesForeignMessageWith404 was vacuous. httpGetShellOutput has a second 404 gate after the service call — `if !ok || message.TaskSessionID != c.Param("id")` — and ExtractShellExecOutput returns ok=false for nil metadata. The fixture had none, so the route answered 404 through that gate whether or not scoping denied the read. Confirmed by mutation rather than inspection: with the authorization removed from GetMessage the test still passed. With valid shell-output metadata in the fixture it returns 200 and the victim's stdout, so the test fails as intended; restoring the guard makes it pass again. Also mutation-tested the sibling denial tests, since this is a defect class rather than a one-off: - TestHTTPGetTaskSessionDeniesForeignSessionWith404 — sound, returns 200 with the session DTO when the guard is removed. - TestProcessRoutesDenyForeignSession — catches removal, though via the documented nil-lifecycleMgr panic rather than a status assertion. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * test(auth): add an in-test witness to the shell-output scoping test Review suggestion. The fixture fix made the test exercise the guard, but nothing in the test proved the 404 came from authorization rather than the route's post-service `!ok` gate — which is exactly how it went vacuous in the first place. The evidence lived in a commit message describing a manual mutation run, not in the test. Now asserts both directions, matching the approve-route test: the non-owner gets 404, and the owner does NOT. If the assertion ever stops proving authorization, the owner case fails and says so. Re-verified by mutation: removing the guard from GetMessage fails the test with 200 and the victim's stdout; restoring it passes. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Kandev Agent <agent@kandev.dev> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Kandev
Manage and run tasks in parallel. Orchestrate agents. Review changes. Ship value.
Features | Workflows | Run as a Service | Debug Logs | Roadmap | Contributing | Architecture | Discord
What
Organize work across kanban and pipeline views with opinionated workflows and execute multiple tasks in parallel. Assign agents from any provider, and review their output in an integrated workspace - file editor, file tree, terminal, browser preview, and git changes in one place. Terminal agent TUIs are great for running agents, but reviewing and iterating on changes there doesn't scale.
Run it locally or self-host it on your own infrastructure and access it from anywhere via Tailscale or any VPN.
Open source, multi-provider, no telemetry, not tied to any cloud.
Vision
Humans stay in control. Define tasks, build agentic workflows with gates, review every change, decide what ships.
- Review-first - Humans support production systems. We need to understand (yet) and trust the code that gets deployed.
- Your workflow - Every team is different, and not every developer uses AI the same way. Define workflows once, share them across the team, and give everyone a consistent process for working with agents - regardless of experience level.
- Remote agents - Running multiple agents on a large codebase can quickly saturate a local machine. The goal is a single control plane: offload execution to servers, orchestrate from anywhere, including your phone.
Features
- Multi-agent support - Claude Code, Codex, GitHub Copilot, Gemini CLI, Amp, Auggie, OpenCode, Cursor, Devin, Qwen, Factory Droid, iFlow, Kilocode, Pi, Kimi, AWS Kiro, Qoder, Trae, Oh My Pi, Grok
- Parallel task execution – start and manage multiple tasks from different sources simultaneously, boosting productivity with AI agents
- Integrated workspace - Built-in terminal, code editor with LSP, git changes panel, embedded vscode and chat in one IDE-like view
- Kanban task management - Drag-and-drop boards, columns, and workflow automation
- Agentic workflows - Multi-step pipelines that mix-and-match agents per step - for example, Claude Code Opus to design a plan, GitHub Copilot Sonnet to implement it, and Codex GPT 5.4 to review the changes. See docs/workflow-tips.md
- Sub-tasks - Agents can spawn sub-tasks that resume from the parent task's session. Useful for splitting a task that has grown too big, or producing several PRs from the same starting point.
- CLI passthrough - Drop into raw agent CLI mode for direct terminal interaction with agents that support it, leveraging their full native TUI
- Workspace isolation - Git worktrees prevent concurrent agents from conflicting
- Multi-repository tasks - Span a single task across multiple repositories, with one worktree per repo, per-repo branches, per-repo PRs, and per-repo grouping in the Changes panel and review dialog
- Flexible runtimes - Run agents as local processes, in isolated Docker containers, on remote servers via SSH, or in cloud executors like sprites.dev
- Runtime settings - Executor profiles, secrets, custom prompts, utility agents, voice mode, and resource metrics are configurable from Settings
- Task-agent MCP - Agents can create subtasks, target sibling repos, attach extra branches for multiple PRs, message other tasks, read conversations, and inspect related tasks
- External MCP - Manage Kandev from outside coding agents over streamable HTTP or SSE, with copyable config snippets for popular agent CLIs
- Workflow portability - Export and import workflows as portable YAML across workspaces or Kandev installs
- Session management - Resume and review agent conversations
- Shareable task snapshots - Publish redacted task conversation snapshots as secret GitHub Gists, with preview and revoke controls
- Stats - Track your productivity with stats on the completed tasks, agent turns, etc
See docs/features.md for the full feature inventory, including settings, secrets, custom prompts, and MCP task capabilities.
In progress: Office mode
We're working on Office mode, a feature-flagged autonomy layer for persistent agent teams. The direction is agent instances with roles and permissions, dashboards, inbox/approvals, routines, task delegation, skills, memory, cost tracking, budgets, and workspace config sync. We'll document Office as a supported feature after it is live.
Integrations
Connect Kandev to the tools your team already uses — pull issues into the kanban, link tasks to PRs, and surface review activity inline.
Supported ACP Agents
| Agent | Launch |
|---|---|
| Claude Code | npx -y @agentclientprotocol/claude-agent-acp@0.61.0 |
| Codex | npx -y @agentclientprotocol/codex-acp@1.1.5 |
| GitHub Copilot | npx -y @github/copilot --acp |
| Gemini CLI | npx -y @google/gemini-cli --acp |
| Amp | npx -y amp-acp |
| Auggie | npx -y @augmentcode/auggie --acp |
| OpenCode | opencode acp |
| Cursor | cursor-agent acp (requires Cursor Pro) |
| Devin | devin acp (install Devin CLI from Devin Desktop or standalone installer) |
| Qwen | npx -y @qwen-code/qwen-code --acp |
| Factory Droid | npx -y droid exec --output-format acp |
| iFlow (beta) | npx -y @iflow-ai/iflow-cli --experimental-acp |
| Kilocode | npx -y @kilocode/cli acp |
| Pi | npx -y pi-acp |
| Kimi | kimi acp (install Kimi CLI from Moonshot AI) |
| Kiro | kiro-cli-chat acp (install Kiro CLI from AWS) |
| Qoder | qodercli --acp (install Qoder CLI) |
| Trae | traecli acp serve (install Trae IDE CLI) |
| Oh My Pi | omp acp (install bun install -g @oh-my-pi/pi-coding-agent) |
| Grok | grok --no-auto-update agent stdio (install npm install -g @xai-official/grok) |
All agents communicate via ACP (Agent Client Protocol). Some agents support ACP natively, while others use ACP adapter packages that bridge their native protocols. CLI Passthrough mode is available when an integration provides a passthrough command. If your agent isn't supported yet, open an issue or submit a PR with the integration. See Adding a New Agent CLI for a step-by-step guide.
Bring your own TUI agents
Kandev can run any agent CLI as a TUI inside a terminal, even when it does not speak ACP. Add the command in agent profile settings and the task starts that agent inside a PTY terminal. CLI passthrough keeps the agent's native terminal UX available while still giving Kandev task tracking, worktrees, review, and session management.
Supported Executors
| Executor | Description |
|---|---|
| Local Process | Runs the agent as a local process on the host machine |
| Docker | Runs the agent in an isolated Docker container |
| SSH | Runs the agent on a remote server over SSH |
| Sprites | Runs the agent in a remote cloud environment via sprites.dev |
Executors support profiles for reusable runtime configuration: prepare scripts, environment variables, credentials, and settings. Worktree-based tasks can also attach multiple repositories or multiple branches from the same repository, which lets one task produce several PRs when the work needs it.
Quick Start
Homebrew (macOS, Linux)
brew install kdlbs/kandev/kandev
kandev
NPX
npx kandev@latest
Requires npm 7+. Installs the CLI plus a platform-matched runtime bundle in one step.
NPM (global)
npm install -g kandev@latest
kandev
Updates
brew upgrade kandevnpx kandev@latest(always uses the latest published version)npm install -g kandev@latest
The worktrees and sqlite db live in ~/.kandev by default. Works on macOS, Linux, and Windows (WSL or native).
From Source
# Clone the repository
git clone git@github.com:kdlbs/kandev.git
cd kandev
# Start in production mode
make start
Prerequisites: Go 1.26.0, Node.js 24, and pnpm 9.15.9 (pinned in mise.toml). Docker is optional and needed only for container runtimes.
High level architecture
graph LR
UI[Web UI] --> Backend
Backend["Backend (Orchestrator)"]
Backend --> W1 & W2 & W3 & W4
subgraph W1[Local Process]
Agent1[Agent CLI] --- WT1[Worktree]
end
subgraph W2[Docker Container]
Agent2[Agent CLI] --- WT2[Worktree]
end
subgraph W3[SSH Remote]
Agent3[Agent CLI] --- WT3[Worktree]
end
subgraph W4[Sprites Cloud]
Agent4[Agent CLI] --- WT4[Worktree]
end
We also want to add support for this remote runtime:
- K8s operator - run agents in a Kubernetes cluster, with auto-scaling and resource management.
Development
Project Structure
apps/
├── backend/ # Go backend (orchestrator, lifecycle, agentctl, WS gateway)
├── web/ # Vite/React frontend (SPA, Zustand, real-time subscriptions)
├── cli/ # CLI tool (npx kandev launcher)
├── desktop/ # Tauri desktop shell
└── packages/ # Shared UI components & types
Prerequisites
- mise for the pinned toolchain in
mise.toml(make bootstrapinstalls it when missing) - Go 1.26.0, Node.js 24, pnpm 9.15.9 (installed by
make bootstrap) - Docker (optional)
Environment Setup
# Install mise tools, workspace dependencies, and git hooks
make bootstrap
# Also install Playwright Chromium browser/deps for E2E
make bootstrap-e2e
Running Dev Servers
# Start everything (backend + frontend with auto ports)
make dev
# Or run separately
make dev-backend # Backend on :38429
make dev-web # Frontend on :37429
Testing & Linting
make test # Run all tests (backend + web)
make lint # Run all linters
make typecheck # TypeScript type checking
make fmt # Format all code
Pre-commit Hooks
# Installed and wired by make bootstrap
make doctor
Comparison to Other Tools
There are a few similar tools in this space, and new ones appearing everyday. Here's what sets this one apart:
- Server-first architecture - The core app runs as a server you can access from any device, including your phone.
- Remote runtimes - Run agents on remote servers via SSH, Docker hosts, and cloud environments, not just your local machine.
- Multi-provider - Use Claude Code, Codex, Copilot, Gemini, Amp, Auggie, OpenCode, Cursor, Devin, Qwen, Droid, iFlow, Kilocode, Pi, Kimi, Kiro, Qoder, Trae, Oh My Pi, and Grok side by side. Not locked to one vendor.
- CLI passthrough and chat - Interact with agents through structured chat messages or, where supported, drop into raw CLI mode for full agent TUI capabilities.
- Open source and self-hostable - No vendor lock-in, no telemetry, runs on your infrastructure.
Contributing
Contributions are welcome! Please read CONTRIBUTING.md before opening a PR.
See the issue tracker for open tasks, or join our Discord to chat with maintainers and other contributors.
Acknowledgments
Built with these excellent open-source projects:
Monaco Editor · Tiptap · xterm.js · dockview · CodeMirror · dnd-kit · Mermaid · Recharts · TanStack Table · Zustand · Shadcn/UI · Radix UI · Tailwind CSS
