Agent teams support

[rubixhacker]

Summary by Sourcery

Add experimental Agent Teams support to the Conductor implement flow, introducing a team-based execution mode alongside the existing sequential implementation mode.

New Features:

• Introduce an experimental Agent Team implementation mode that orchestrates multiple specialized agents in parallel for track execution.
• Add a --team flag to /conductor:implement to trigger Agent Team mode when the Agent Teams feature is enabled via environment variable.
• Define Conductor's role as Team Lead for Agent Teams, including task assignment, coordination, and verification against the track spec.

Enhancements:

• Clarify the standard implementation flow and explicitly route both team and standard modes to the documentation synchronization step.
• Update CLI command documentation to describe the new Agent Teams capability and usage prerequisites.

Documentation:

• Document the experimental Agent Teams feature, including prerequisites, usage of /conductor:implement --team, and Conductor's coordination behavior.

[sourcery-ai]

Reviewer's Guide

Adds an experimental Agent Teams execution mode to the /conductor:implement command, including branching logic on a --team flag, detailed team orchestration protocol, documentation for enabling and using Agent Teams, and minor updates to the standard implementation flow and command description.

Sequence diagram for Agent Teams orchestration in implementation mode

sequenceDiagram
  actor User
  participant Conductor
  participant TracksFile as tracks_md
  participant TrackContext as track_files
  participant AgentPlatform as agent_teams

  User->>Conductor: /conductor:implement --team
  Conductor->>Conductor: Check --team flag
  Conductor->>Conductor: Verify CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS
  Conductor-->>User: Announce starting Agent Team implementation

  Conductor->>TracksFile: Update track status to [~]

  Conductor->>TrackContext: Read plan_md
  Conductor->>TrackContext: Read spec_md
  Conductor->>TrackContext: Read tech_stack_md
  Conductor->>TrackContext: Read workflow_md

  Conductor->>Conductor: Analyze plan_md to determine roles

  Conductor->>AgentPlatform: Spawn agent team with roles and responsibilities
  AgentPlatform-->>Conductor: Teammates ready

  loop Manage execution
    Conductor->>AgentPlatform: Assign tasks to teammates
    AgentPlatform-->>Conductor: Teammate outputs
    Conductor->>Conductor: Review outputs and resolve conflicts
    Conductor->>Conductor: Verify against spec_md
  end

  Conductor->>TracksFile: Update track status to [x]
  Conductor-->>User: Announce completion
  Conductor->>Conductor: Proceed to 6_0_SYNCHRONIZE_PROJECT_DOCUMENTATION

Loading

File-Level Changes

Change	Details	Files
Introduce Implementation Mode branching with support for experimental Agent Teams in the implement command protocol.	Rename the main implementation section from 'Track Implementation' to 'Implementation Mode' and describe how to detect the --team flag. Define branching behavior based on the presence of the --team flag, including environment variable gating via CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS for team mode. Specify failure behavior and messaging when the --team flag is used but Agent Teams are not enabled, and fall-through to standard implementation when not used.	commands/implement.md
Define a detailed Team Orchestration workflow for Agent Teams, including team formation, coordination, and completion steps.	Add a 'Team Orchestration (Experimental)' section describing how to announce team usage, load track context, and update track status. Document how to analyze the track plan to derive roles and spawn an agent team using natural language prompting, including an example prompt template. Describe responsibilities of the Team Lead agent for assigning tasks, reviewing outputs, resolving conflicts, and ensuring adherence to spec and project docs, plus finalization steps that update track status and sync documentation.	commands/implement.md
Clarify Standard Implementation flow and ensure it routes into project documentation synchronization.	Introduce a 'Standard Implementation' section explicitly labeled as sequential execution while preserving existing step details. Append a final step to proceed to '6.0 SYNCHRONIZE PROJECT DOCUMENTATION' after marking a track complete in the standard flow.	commands/implement.md
Document Agent Teams usage and prerequisites in the main CLAUDE documentation and surface the --team flag in the implement command description.	Update the /conductor:implement command description to mention use of the --team flag for Agent Teams. Add an 'Agent Teams (Experimental)' section describing the feature conceptually, including environment variable prerequisite and usage instructions for running /conductor:implement --team and Conductor acting as Team Lead.	CLAUDE.md

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey - I've left some high level feedback:

• Consider briefly clarifying what happens when /conductor:implement --team is used but CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS is not set (e.g., error messaging/behavior) in CLAUDE.md so users don’t have to infer it from implement.md alone.
• The instructions for Agent Teams (env var requirement, --team flag usage, team lead role) are now split between implement.md and CLAUDE.md; consider consolidating or cross-linking key details to avoid divergence or duplication over time.
• In the TEAM ORCHESTRATION section, you might want to explicitly define what “spawn the agent team using your natural language capabilities” means in operational terms (e.g., expected command or interaction pattern) to reduce ambiguity for future maintainers and tools interpreting the protocol.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- Consider briefly clarifying what happens when `/conductor:implement --team` is used but `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS` is not set (e.g., error messaging/behavior) in CLAUDE.md so users don’t have to infer it from implement.md alone.
- The instructions for Agent Teams (env var requirement, `--team` flag usage, team lead role) are now split between implement.md and CLAUDE.md; consider consolidating or cross-linking key details to avoid divergence or duplication over time.
- In the TEAM ORCHESTRATION section, you might want to explicitly define what “spawn the agent team using your natural language capabilities” means in operational terms (e.g., expected command or interaction pattern) to reduce ambiguity for future maintainers and tools interpreting the protocol.

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}