More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    claude agent team companion

    1

    - Five reusable templates: specification, ticket, handoff, decision record and acceptance report.

    $20

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    claude agent team companion

    claude agent team companion

    Example session with this skill installed

    Request

    “Use Claude Agent Team Companion to implement an issue-filter helper. Support optional status and assignee filters, combine them with AND, preserve input order, and verify the behavior.”

    Worked result

    This is the package’s synthetic teaching exercise. Its Python acceptance checks were executed successfully before packaging. It is not a customer project or a measured multi-agent run.

    The workflow establishes a concrete contract:

    • No filters returns all issues.
    • Status and assignee filters combine with AND.
    • Matching is case-sensitive.
    • Input order is preserved and the source list is not modified.
    • Invalid filter values raise ValueError.

    For this small helper, the workflow chooses one agent rather than creating unnecessary coordination.

    Given

    IDStatusAssignee
    1openAda
    2closedAda
    3openLin

    The verified results include

    FiltersReturned IDs
    None1, 2, 3
    Status: open1, 3
    Assignee: Ada1, 2
    Status: open; assignee: Ada1
    Status: closed; assignee: LinEmpty

    Executed exercise output

    feature: acceptance checks passed
    

    Visual collaboration preview

    The accompanying dashboard screenshots use clearly labeled synthetic participants and messages to illustrate dispatch, peer communication, handoff and reporting.

    In a supported live environment, characters represent actual participants and communication comes from available runtime records. The board does not create agents or prove that a task passed its acceptance checks.

    • Read your context and instructions
    • Compiled the claude agent team

    Issue-filter helper: worked result

    This is a synthetic teaching exercise included in the package. Its Python acceptance checks were executed successfully before packaging. It is not a customer project or a measured multi-agent run.

    Agreed behavior

    • Optional status and assignee filters combine with AND.
    • No filters returns all issues.
    • Matching is case-sensitive.
    • Input order is preserved.
    • The source list is not modified.
    • Invalid filter values raise ValueError.

    Implementation

    def filter_issues(issues, status=None, assignee=None):
        if status is not None and status not in ("open", "closed"):
            raise ValueError("status must be open, closed, or None")
    
        if assignee is not None and (
            not isinstance(assignee, str) or not assignee
        ):
            raise ValueError("assignee must be a nonempty string or None")
    
        return [
            row for row in issues
            if (status is None or row["status"] == status)
            and (assignee is None or row["assignee"] == assignee)
        ]
    

    Verified example

    Input issues

    IDStatusAssignee
    1openAda
    2closedAda
    3openLin

    Acceptance results

    FiltersReturned IDs
    None1, 2, 3
    Status: open1, 3
    Assignee: Ada1, 2
    Status: open; assignee: Ada1
    Status: closed; assignee: LinEmpty

    Additional checks covered empty input, invalid filter values, case-sensitive matching, unchanged source data and returned object identity.

    Executed exercise output

    feature: acceptance checks passed
    

    Coordination decision and handoff

    This helper is a small, bounded change, so the workflow uses one agent. It does not create unnecessary workers.

    Exported interface: filter_issues(issues, status=None, assignee=None).

    For integration into a real application, the next owner should verify that UI fields map to the correct arguments and that an empty result displays correctly. Those UI checks are outside this exercise.

    The separately installed dashboard can visualize supported real participants and recorded communication. Listing screenshots use explicitly labeled synthetic data to illustrate dispatch, peer messages, handoffs and reports.

    Next steps

    1. Initialize the .team/ directory and commit the spec.md file.
    2. Assign ticket_001.md to a subagent worker if delegation is enabled.
    3. Verify the final implementation using the npm test command provided in the acceptance template.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Make agent collaboration easier to follow

    Guide your coding agent through feature delivery, bug investigation and code review—with clear ownership, useful handoffs and verifiable results.

    Pair the skill with the free Claude Agent Team local board to see supported participants as animated characters, follow model and task activity, and inspect recorded communication between agents.

    What you get

    • Three guided workflows: feature delivery, bug investigation and code review.
    • Five reusable templates: specification, ticket, handoff, decision record and acceptance report.
    • Three executable Python exercises with inputs, code and expected results.
    • Team-sizing guidance that respects your tools and budget.
    • A visual-collaboration playbook for connecting real participants, messages and handoffs to the local board.

    How it works

    Describe your task to your connected coding agent. The skill helps it inspect the project, define acceptance conditions, assign file ownership, coordinate permitted work and verify the outcome.

    Small tasks stay single-agent. Larger tasks use delegation only when your environment supports it and you allow it.

    Requirements and boundaries

    The purchase provides workflow instructions and supporting text. The free dashboard is installed separately from GitHub.

    Live monitoring requires local session files and suitable local tools. Supported Claude Code and Codex activity can be visualized; Codex monitoring is read-only. Recorded send calls are not read receipts, and missing communication cannot be reconstructed.

    Model access, API usage and project dependencies are separate. No hosted dashboard or guaranteed productivity improvement is included.

    Free project: https://github.com/another-bbzj/claude-agent-team

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.

    Requires local shell/file access to execute plans. Does not include the dashboard installer or model-calling capabilities. Works best with the Claude Agent Team repository structure.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 16 days ago

    Software
    Python 3.8+
    Optional
    Claude CodeGit

    Claude Code is preferred. Python runs the optional local dashboard and the Python exercises. Git clones the free dashboard.

    Permissions

    Read Files
    Write Files
    Terminal / Shell
    Network Access

    File access is limited to the user's authorized project. The skill reads relevant code and instructions, creates project-local tickets and handoffs, and may modify implementation or test files when requested. Shell access is used to run project checks and optional setup commands. Network access is used only when requested to obtain the free dashboard or communicate with its local loopback service. The optional dashboard is installed separately and reads supported local Claude Code or Codex session files. Installation and access to session data require the user's authorization. Codex monitoring is read-only. The skill does not require additional connected service accounts, collect credentials, or upload project files or session logs to an external service. Planning-only use does not require local execution.

    Listed16 days ago

    What's inside

    Frequently Asked Questions

    It prevents 'conflicting writes' and 'interface drift' by using a structured ticketing system that assigns specific file ownership and interface contracts to each sub-agent before they begin work.

    The skill requires an environment with local file system access and shell execution capabilities, specifically designed to work within the Claude Agent Team framework.

    No, this skill provides the logical framework, templates, and workflow protocols; the visual dashboard itself is a separate component of the Claude Agent Team project.

    Yes, it includes specialized workflows for bug investigation and code reviews, enabling agents to isolate failures with reproducible test cases and assess edge cases in proposed PRs.

    While optimized for Claude-based workflows, the framework uses language-agnostic interface contracts and shell-based testing, making it compatible with any codebase that supports standard CLI tools.

    New in Agents & Orchestration