Coding agent

Zoo Code

Roo successor for controlled, model-agnostic coding inside VS Code.

Status
Active
License
Apache-2.0
Evidence
Documented + code-verifiable, 27 sources
Product record checked
2026-07-27

At a glance

An Apache-licensed VS Code agent continuing the Roo Code lineage, with broad provider and subscription access, local models, custom modes, hierarchical subtasks, MCP, semantic indexing, worktree workflows, task history, and shadow-Git file checkpoints.

Good choice if

  • Vibe coders who want proposed tools and diffs visible in VS Code before they run
  • Roo Code users who need a maintained migration path with compatible settings and task history
  • Model-agnostic IDE workflows using modes, MCP, local models, semantic indexing, subtasks, and worktrees

Check before choosing

  • Commands and file tools run with the local user's host permissions; worktrees and .rooignore scope work but are not an OS or container sandbox
  • Extension actions ask by default, but auto-approval can be enabled and sending a queued message implicitly approves the next pending tool, file, or command action even when auto-approval is disabled
  • Shadow-Git checkpoints restore task-scoped file state and are enabled by default, but they are not created before commands and cannot reverse external command side effects
See 4 more considerations
  • The new_task tool suspends the parent and opens one child as the sole active task; simultaneous work instead uses separate user-managed VS Code windows and Git worktrees, not an automatically parallel delegated swarm
  • The public website says Zoo works headless, but the inspected CLI package is private, still uses Roo names and upstream Roo install URLs, and has no Zoo-owned CLI release asset; it is not counted as a dependable public CI interface
  • No built-in browser controller was found in the tool registry; web access depends on a browsing-capable provider or an external Playwright or Puppeteer MCP server
  • The repository has extensive engineering and end-to-end tests but no coding-harness evaluation suite, independent benchmark, or direct scientific paper; inherited release-note model scores are not Zoo Code harness evidence

Why it qualifies as a coding harness

This confirms category fit, not product quality. Every required criterion links back to first-party evidence.

Qualifies4 of 4 required criteria evidenced
  • Adaptive agent loop

    Documented

    The system repeatedly observes results and chooses the next action instead of following a fixed one-pass graph.

  • Repository tool execution

    Documented

    The system can use tools to inspect and change a repository or its execution environment.

  • Task-aware context management

    Documented

    The runtime assembles, updates, compacts, retrieves, or persists task-relevant context while work proceeds.

  • Model-independent runtime control

    Documented

    Permissions, budgets, interruption, policy, or stop controls operate outside the model's own text generation.

Membership establishes category fit only. It does not score quality, safety, autonomy, model capability, or benchmark performance. · Read the membership rule.

How it works under the hood

Seven mechanisms mapped from first-party records. These labels describe what the harness provides, not how intelligent its model is.

7/7layers documented
  • Execution & isolationWorkspace isolationDocumented mechanism, not a performance score.
  • Tooling & integrationsExtensible toolsDocumented mechanism, not a performance score.
  • Context & stateManaged contextDocumented mechanism, not a performance score.
  • Lifecycle & recoveryCheckpoint/rewindDocumented mechanism, not a performance score.
  • ObservabilityLogs/transcriptsDocumented mechanism, not a performance score.
  • VerificationTool-assistedDocumented mechanism, not a performance score.
  • Governance & permissionsApproval promptsDocumented mechanism, not a performance score.

Public code audit

4/5public artifacts present
Security policy
Present at inspected commit
CI workflow
Present at inspected commit
Automated tests
Present at inspected commit
Evaluation assets
Not found
Contributor documentation
Present at inspected commit

The repository contains 754 unit, integration, and end-to-end test files, but no coding-harness evaluation suite or independent benchmark. CONTRIBUTING describes evals as a future goal; the source CLI remains private and Roo-branded despite a public headless claim.

Inspect commit d27153a251d2051b6a8e73d305b06ffbc5ac6970, checked 2026-07-27

Measured configurations

No benchmark run passes the full metadata admission policy for this harness yet. Missing data is not scored as zero.

Benchmark policy and all runs

Capability support

Documented first-class product support, checked against the sources below.

External tools (MCP)
DocumentedProduct-supported MCP integrationSource · checked 2026-07-27The source establishes the mechanism, not its quality or availability in every mode.
Local models
DocumentedLocal or self-hosted model pathSource · checked 2026-07-27The source establishes the mechanism, not its quality or availability in every mode.
Agent parallelism
DocumentedDelegated or parallel agent workflowSource · checked 2026-07-27The source establishes the mechanism, not its quality or availability in every mode.
Runs without an open UI
Not documentedNo first-class support established by the current recordAbsence of current documentation is not proof that the capability is impossible.
Browser control
Not documentedNo first-class support established by the current recordAbsence of current documentation is not proof that the capability is impossible.
Isolated execution
Not documentedNo first-class support established by the current recordAbsence of current documentation is not proof that the capability is impossible.
Undo file changes
DocumentedProduct-supported file or session rollbackSource · checked 2026-07-27The source establishes the mechanism, not its quality or availability in every mode.

Primary evidence

Each capability claim is tied to a first-party record and a verification date.

Product record checked 2026-07-27
View 19 additional sources
Additional first-party evidence19 sources
Codebase indexingTree-sitter parsing, embedding providers, Qdrant storage, incremental updates, multi-folder behavior, and privacy boundariesOfficial docsCheckpointsDefault-on shadow-Git snapshots, task-scoped file restoration, diffs, and the lack of pre-command checkpointsOfficial docsGit worktreesSeparate branches and VS Code windows for user-managed parallel work, plus worktree requirements and limitationsOfficial docsFile access rules.rooignore enforcement for file tools and its explicit limitation as neither full command coverage nor a system sandboxOfficial docsMessage queueingSequential queued prompts and the important behavior that a queued message implicitly approves the next pending actionOfficial docsDiagnostic exportVersioned error details, full task history, actions, environment, configuration, and support-oriented diagnostic exportOfficial docsRoo to Zoo migrationProject succession, configuration migration, compatibility boundaries, and user transitionOfficial docsZoo Code 3.72.0 releaseLatest stable extension version, publication date, VSIX artifact, provider changes, and subtask fixesOfficial announcementRepository overview at inspected commitActive Roo succession, VS Code extension scope, current feature summary, contribution path, and release lineageOfficial repositoryApache license at inspected commitApache-2.0 terms for the reviewed source repositoryOfficial repositorySecurity policy at inspected commitSupported-version scope, private vulnerability reporting, acknowledgment target, and remediation targetOfficial repositoryContribution guide at inspected commitIssue-first contribution policy, testing requirements, and the stated future goal of establishing harness evaluationsOfficial repositoryCode QA workflow at inspected commitCross-platform unit and integration tests, coverage, lint, type checks, dependency review, and security-oriented source checksOfficial repositoryDelegation implementation at inspected commitApproval-gated hierarchical child creation and replacement of the parent by one sole active child taskOfficial repositoryAuto-approval policy at inspected commitGranular read, write, command, MCP, mode, subtask, and follow-up auto-approval decisionsOfficial repositoryWorktree service at inspected commitCode-level Git worktree creation, listing, switching, deletion, branch handling, and host-process executionOfficial repositorySource CLI readme at inspected commitPrint and streaming JSON modes, source-build instructions, default auto-approval, and stale Roo-owned installation pathsOfficial repositorySource CLI manifest at inspected commitPrivate Roo-branded package metadata and version, limiting its status as a distributed Zoo product interfaceOfficial repositorySource CLI behavior at inspected commitNon-interactive execution, provider selection, persistence, and automatic approval unless approval mode is requestedOfficial repository

Ecosystem discovery

OpenRouter apps Discovery signal only; usage rank is not used as a quality or capability score. Observed 2026-07-27.