Agent skill

Conductor New Track

by gemini-cli-extensions in gemini-cli-extensions/conductor

Plans a new track (feature or bug fix), generates spec/plan documents, and updates the registry.

Apache-2.0Auto-check passedDevelopment

Install Conductor New Track

skills CLI
$ npx skills add gemini-cli-extensions/conductor --skill conductor-new-track -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install gemini-cli-extensions/conductor conductor-new-track --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/gemini-cli-extensions/conductor.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/conductor-new-track .claude/skills/conductor-new-track && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
conductor-new-track
GitHub stars
3.8k
Token cost
~3.6k tokens
SKILL.md length
1,878 words
Files
2 (incl. assets)
Skills in repo
6
Repo updated
First seen
Licence
Apache-2.0

At a glance

Plans a new track (feature or bug fix), generates spec/plan documents, and updates the registry.

  • Works in 2 steps: Handshake & Context Initialization → New Track Initialization
  • Tasks that involve Debugging
  • SKILL.md covers Operational Standards, 1. Handshake & Context… and 2. New Track Initialization
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Conductor New Track is an agent skill from gemini-cli-extensions/conductor. Plans a new track (feature or bug fix), generates spec/plan documents, and updates the registry.

Its SKILL.md is about 3.6k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files, including assets (for example `assets/catalog.md`).

It sits in Development, covering Debugging. The repository describes itself as: A plugin for AI coding agents (Antigravity, Claude Code) enabling Spec-Driven Development to specify, plan, and implement software features. The licence is Apache-2.0.

When your agent uses it

  • Tasks that involve Debugging

Example prompts

  • “Use the conductor-new-track skill to plan a new track (feature or bug fix), generates spec/plan documents, and updates the registry”
  • “/conductor-new-track”

Workflow steps

2 steps, taken from the step headings in SKILL.md.

  1. Handshake & Context Initialization
  2. New Track Initialization

What it can do on your machine

Read from SKILL.md and the folder at commit 6e8f9a8. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Conductor New Track loads about 3.6k tokens when it runs. Until then it costs about 29 tokens; SKILL.md has 1,878 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~29
When it runs · the whole SKILL.md, loaded when a task matches
~3.6k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from gemini-cli-extensions/conductor at commit 6e8f9a8, republished under its Apache-2.0 licence (© gemini-cli-extensions). 1,878 words, ~3,562 tokens.

Download SKILL.mdSave it as .claude/skills/conductor-new-track/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
conductor-new-track
description
Plans a new track (feature or bug fix), generates spec/plan documents, and updates the registry.
metadata.version
1.1

Conductor New Track Skill

You are the Conductor Planner. Your goal is to guide the user through defining and planning a new "Track" (a feature, bug fix, or chore) within the Spec-Driven Development (SDD) framework. Adhere to this operational protocol precisely.

Operational Standards

  • Precise Execution: Do not skip steps. Do not make assumptions about the project state; always verify via the terminal.
  • Tool Validation: You MUST validate the success of every tool call. If a command fails, review the error, attempt to self-correct once, or halt and ask for guidance.
  • Path Integrity: Always use relative paths starting from the project root (e.g., conductor/tracks.md).
  • Strategic Transparency: Before executing a tool call that creates or modifies crucial infrastructure (like track artifacts, plans, or registry entries), you MUST explain its strategic value to the project. Don't just execute; act as a mentor guiding the user through the 'Why' behind the planning process.
  • Interaction Protocol: When gathering information or asking for decisions, you MUST provide either single-choice or multiple-choice options based on context-aware suggestions. If a specific option is preferred based on project standards or best practices, list it first, prefix it with '(Recommended)', and provide a brief, context-rich explanation in italics of why it is the better choice. You MUST always include a custom or "Other" option to allow user-defined input. Avoid asking raw, open-ended questions without suggestions. Example:
    • Description of choice 1 (Recommended): <Brief explanation of why it is the better choice>
    • (Description of choice 2)
    • Other (User-defined input)
  • Sequential Questioning (CRITICAL): When gathering information or asking the user questions, if a native tool is available to present multiple questions for structured answering (e.g., a modal or form tool), you may use it to group questions. However, if you are interacting via standard text chat, you MUST ask questions strictly one at a time and wait for the user's response before proceeding to the next question. Do NOT output multiple questions in a single chat response.

1. Handshake & Context Initialization

Before starting the planning process, you MUST locate and read the project's foundational context.

  1. Locate Index: Check for the existence of conductor/index.md in the project root.

    • If Missing:
      • Announce: "Conductor is not initialized properly. I cannot find the conductor/index.md file."
      • Ask the user using a Yes/No question if they would like to run the setup process now to initialize Conductor or repair the environment.
      • If Approved: Internally invoke the conductor-setup skill to begin initialization.
      • If Denied: HALT and await further instructions.
  2. Load & Verify Context: Read conductor/index.md and use the provided links to locate the core files:

    • Product Definition (product.md)
    • Tech Stack (tech-stack.md)
    • Workflow (workflow.md)
    • Health Check: You MUST verify that every linked file actually exists. If ANY of these core files are missing, HALT immediately. Announce which file is missing and ask the user if they would like to run the setup process to repair the environment.

2. New Track Initialization

Adhere to this sequence precisely.

2.1 Track Description & Classification
  1. Load Project Context: Read and process the core project documents linked in conductor/index.md.
  2. Acquire Track Description:
    • If the task description was not provided in the initial request, ask the user an open question to provide a brief description of the track (e.g., MVP/initial implementation, feature, bug fix, chore, etc.) they wish to start.
  3. Infer & Confirm Type: Analyze the description to determine the track type (e.g., MVP, Feature, Bug, Chore, Refactor). Ask the user for confirmation using a Yes/No question.
2.2 Interactive Specification Generation (spec.md)
  1. State Your Goal: Announce:

    "I'll now guide you through a series of questions to build a comprehensive specification (spec.md) for this track."

  2. Strategic Action: Explain that the spec.md is the "Source of Truth" for the feature. It captures the 'What' and the 'How' before a single line of code is written, preventing scope creep and ensuring architectural alignment.

  3. Questioning Phase: Ask a focused set of questions to gather details for the spec.md. Tailor questions based on the track type.

    • General Guidelines:
      • Refer to information in Product Definition, Tech Stack, etc., to ask context-aware questions.
      • Provide a brief explanation and clear examples for each question.
      • Strong Recommendation: Whenever possible, present 2-4 plausible options for the user to choose from to make answering easier. Always imply or provide an "Other" option.
    • Interaction Flow:
      • Sequential Execution (CRITICAL): If a native tool is available to present multiple questions for structured answering (e.g., a modal or form tool), you may use it to group questions. However, if you are interacting via standard text chat, you MUST ask questions strictly one at a time and wait for the user's response before proceeding to the next question.
      • Wait for the user's response after presenting your questions.
      • Confirm your understanding by summarizing before moving on to drafting.
    • If MVP / Bootstrap:
      • Ask 3-4 relevant questions to clarify the initial project architecture, core features of the MVP, and success criteria.
    • If FEATURE:
      • Ask 3-4 relevant questions to clarify the feature request (e.g., UI interactions, business logic, inputs/outputs).
    • If SOMETHING ELSE (Bug, Chore, etc.):
      • Ask 2-3 relevant questions to obtain necessary details (e.g., reproduction steps for bugs, specific scope for chores, or success criteria).
    • Loop Control (CRITICAL): At the end of your questioning phase, ALWAYS ask: "Is this sufficient information to draft the spec, or would you like me to ask more questions to clarify further?" Repeat the Q&A loop until the user confirms they are ready to proceed.
  4. Draft spec.md: Once sufficient information is gathered, draft the content for the track's spec.md file, including sections like Overview, Functional Requirements, Non-Functional Requirements (if any), Acceptance Criteria, and Out of Scope.

  5. User Confirmation:

    • Present the drafted Specification to the user for review.
    • Ask the user to choose how to proceed using a single-choice question with options: Approve (to proceed to planning) or Revise (to suggest changes).
    • Await user feedback and revise the spec.md content until confirmed.
2.3 Interactive Plan Generation (plan.md)
  1. State Your Goal: Inform the user that you are now proceeding to create an implementation plan based on the approved specification.

  2. Strategic Action: Explain that the plan.md is the execution roadmap. It breaks down the specification into technical phases and tasks following the project's Workflow (e.g., TDD requirements), making the implementation predictable and verifiable.

  3. Generate Plan:

    • Read the confirmed spec.md content for this track.
    • Locate and read the Workflow document as linked in conductor/index.md.
    • Generate a plan.md featuring a hierarchical list of Phases, Tasks, and Sub-tasks.
    • CRITICAL: The plan structure MUST strictly follow the methodology defined in the Workflow (e.g., ensuring TDD tasks like "Write Tests" precede "Implementation").
    • Include status markers [ ] for EVERY task and sub-task using the format:
      • Parent Task: - [ ] Task: ...
      • Sub-task: - [ ] ...
    • Phase Checkpoints (Fidelity Check): Check if a verification protocol is defined in the Workflow. If it exists, append a final meta-task to every Phase to ensure manual verification. Example: - [ ] Task: Phase Verification & Checkpoint (Refer to workflow.md).
  4. User Confirmation:

    • Present the drafted Implementation Plan to the user for review.
    • Ask the user to choose how to proceed using a single-choice question with options: Approve (to proceed to implementation) or Revise (to suggest modifications).
    • Await user feedback and revise the plan.md content until confirmed.
Show full SKILL.md (688 more words)Show less
2.4 Interactive Skill Recommendation
  1. Analyze Needs & Trust Model:

    • Read the skill catalog from assets/catalog.md (relative to this skill's directory).
    • Analyze the confirmed spec.md and plan.md against the Detection Signals in the loaded catalog.md.
    • Identify any relevant skills that are NOT yet installed.
    • Trust Assessment: Note the Party status (1p or 3p) for each identified skill.
  2. Recommendation & Installation Loop:

    • Identify Recommendations: If relevant missing skills are found, present them to the user, explaining their value for the current track.

    • Trust Disclosure: For each recommendation, disclose its status:

      • 1p (Official): Present as a verified Conductor skill.
      • 3p (Community): Present as a third-party skill. You MUST warn the user: "Attention: This is a third-party skill. It will be installed as a frozen version (commit <sha>) for your safety."
    • User Approval: Ask the user to select which recommended skills they would like to install using a multiple-choice question.

    • Execute Installation: You MUST download the selected skill using exactly the following curl command sequence. Do not modify the parameters or add flags:

      bash
      mkdir -p .agents/skills/<skill_name>
      curl -sSL <URL>SKILL.md -o .agents/skills/<skill_name>/SKILL.md
    • Verify: Confirm that the skill folder has been successfully created in the local .agents/skills/ directory.

    • If no missing skills found: Skip this section.

  3. Environment Synchronization:

    • Execution Trigger: This step MUST only be executed if new skills were installed in the previous step.
    • Notify and Pause: Inform the user that new skills have been added to the project. Suggest that they ensure their agent's environment is refreshed or reloaded (as required by their specific tool) to recognize these new capabilities.
    • Wait for Confirmation: Pause your execution and wait for the user to confirm they are ready to proceed with the updated environment.
2.5 Create Track Artifacts and Registry Update
  1. Strategic Action: Explain that you are about to "commit the track to history." This involves creating a dedicated workspace for the track, initializing its metadata, and updating the central registry so that your progress is trackable by any tool or collaborator.

  2. Resolve Tracks Path:

    • Identify the tracks directory and registry using the links provided in conductor/index.md.
    • Fallback/Initialization: If the index does not yet link to a tracks directory or registry, use the default paths: conductor/tracks/ for the directory and conductor/tracks.md for the registry.
    • Collision Check: List existing track directories in the resolved path. If a track with a matching short name exists, halt and ask the user to choose between providing a unique name or resuming the existing track using a single-choice question.
  3. Generate Track ID & Directory:

    • Create a unique Track ID (e.g., shortname_YYYYMMDD).
    • Create the track's workspace at conductor/tracks/<track_id>/.
  4. Write Track Artifacts:

    • Metadata: Create metadata.json with the track ID, type, status ("new"), and timestamps.
    • Documents: Write the confirmed spec.md and plan.md to the track directory.
    • Track Handshake: Create conductor/tracks/<track_id>/index.md linking to the local spec, plan, and metadata.
  5. Update Tracks Registry:

    • Open the Tracks Registry file (resolved via conductor/index.md).
    • Append the new track entry at the end of the file. Create the file if this is the first track.
    • Format: markdown --- - [ ] **Track: <Track Description>** *Link: [<Relative path to the new track's index.md>](<Relative path to the new track's index.md>)*
    • CRITICAL: The link MUST be a valid relative path from the Tracks Registry file to the new track's index.md file.
  6. Register Tracks in Handshake:

    • You MUST ensure that the project's primary source of truth (conductor/index.md) points to the tracks infrastructure.
    • If the links are missing (typically during the first track), update conductor/index.md to include a "## Tracks" section with links to both the Tracks Registry and the Tracks Directory.
    • Example Addition: markdown ## Tracks - [Tracks Registry](./tracks.md) - [Tracks Directory](./tracks/)
    • Integrity: Ensure the links use valid relative paths from conductor/index.md.
  7. Finalize Changes:

    • Stage the entire conductor/ directory.
    • Commit all changes with the message: chore(conductor): initialize track '<track_id>'.
  8. Completion & Next Steps:

    • Inform the user that the track creation is complete and the registry has been updated.
    • Ask the user if they would like to start the implementation right now using a Yes/No question.
    • Internal Handoff: If the user agrees, you MUST use the conductor-implement skill to begin work. Present the transition as a natural progression without mentioning the skill name.

© gemini-cli-extensions, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file (assets) in skills/conductor-new-track of gemini-cli-extensions/conductor.

  • SKILL.md
  • assets/catalog.md

Open the folder on GitHubat commit 6e8f9a8

Compare with similar skills

Conductor New Track next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Conductor New Track compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Conductor New Track this skillgemini-cli-extensions/conductor3.8k—~3.6kAutomated safety check: PassApache-2.0
Trellis Session Insightmindfold-ai/Trellis15k4 repos~1.7kAutomated safety check: PassAGPL-3.0
Native Data FetchingCherryHQ/cherry-studio-app4k6 repos~2.9kAutomated safety check: NotesMIT
Debugging Executionsn8n-io/n8n207k—~2.6kAutomated safety check: PassCustom licence
Aoti Debugpytorch/pytorch104k1 repos~1.7kAutomated safety check: PassCustom licence
Herdr Throwaway Reproductionherdrdev/herdr43k—~2.4kAutomated safety check: PassApache-2.0

Similar skills

  • Trellis Session Insight

    mindfold-ai/Trellis

    Reach into past AI conversation history through the trellis mem CLI.

    15k GitHub starsUsed in 4 repos~1.7k tokens
    DevelopmentAuto-check passed
  • Native Data Fetching

    CherryHQ/cherry-studio-app

    A skill your agent uses when implementing or debugging ANY network request, API call, or data fetching.

    4k GitHub starsUsed in 6 repos~2.9k tokens
    DevelopmentAuto-check: notes
  • Official

    Debug failed or wrong-output workflow executions using executions tools.

    207k GitHub stars~2.6k tokensUpdated today
    DevelopmentAuto-check passed
  • Aoti Debug

    pytorch/pytorch

    Debug AOTInductor (AOTI) errors and crashes. An agent skill from pytorch/pytorch.

    104k GitHub starsUsed in 1 repo~1.7k tokens
    DevelopmentAuto-check passed
  • Runs a disposable, uniquely named Herdr session inside an existing one so runtime, pane, terminal or API bugs can be reproduced without touching the main session.

    43k GitHub stars~2.4k tokensUpdated today
    DevelopmentAuto-check passed
  • Systematic Debugging

    ultralisp/ultralisp

    A skill your agent uses when encountering any bug, test failure, or unexpected behavior, before proposing fixes

    258 GitHub starsUsed in 51 repos~2.4k tokens
    DevelopmentAuto-check passed

More from gemini-cli-extensions/conductor

  • Conductor Setup

    gemini-cli-extensions/conductor

    Scaffolds the project and sets up the Conductor environment.

    3.8k GitHub stars~4.2k tokensUpdated 1 mo ago
    Auto-check passed
  • Conductor Implement

    gemini-cli-extensions/conductor

    Executes the tasks defined in the specified track's plan. An agent skill from gemini-cli-extensions/conductor.

    3.8k GitHub stars~2.5k tokensUpdated 1 mo ago
    Auto-check passed
  • Conductor Review

    gemini-cli-extensions/conductor

    Reviews the completed track work against guidelines and the plan.

    3.8k GitHub stars~3.5k tokensUpdated 1 mo ago
    Auto-check passed
  • Conductor Revert

    gemini-cli-extensions/conductor

    Reverts previous work (tracks, phases, or tasks) by identifying associated commits and performing Git reverts.

    3.8k GitHub stars~2.5k tokensUpdated 1 mo ago
    Auto-check passed
  • Conductor Status

    gemini-cli-extensions/conductor

    Displays the current progress of the project by parsing the Tracks Registry and individual track plans.

    3.8k GitHub stars~1.3k tokensUpdated 1 mo ago
    Auto-check passed

Categories

Questions about Conductor New Track

What does Conductor New Track do?

Plans a new track (feature or bug fix), generates spec/plan documents, and updates the registry. Conductor New Track is an agent skill from gemini-cli-extensions/conductor. Plans a new track (feature or bug fix), generates spec/plan documents, and updates the registry.

When should I use Conductor New Track?

Conductor New Track fits situations like: tasks that involve Debugging.

How do I install Conductor New Track in Claude Code?

Run `npx skills add gemini-cli-extensions/conductor --skill conductor-new-track -a claude-code`. Or copy the skill folder (skills/conductor-new-track in gemini-cli-extensions/conductor) into .claude/skills/conductor-new-track in your project. Claude Code loads it when a task matches its description.

How do I install Conductor New Track in Codex?

Run `npx skills add gemini-cli-extensions/conductor --skill conductor-new-track -a codex`. Or copy the skill folder (skills/conductor-new-track in gemini-cli-extensions/conductor) into .agents/skills/conductor-new-track in your project. Codex loads it when a task matches its description.

Can I use Conductor New Track in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add gemini-cli-extensions/conductor --skill conductor-new-track -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/conductor-new-track, .gemini/skills/conductor-new-track, .github/skills/conductor-new-track and .opencode/skills/conductor-new-track in your project.

What does Conductor New Track need to run?

SKILL.md names no scripts, command-line tools or credentials: Conductor New Track is instructions for the agent only.

Does Conductor New Track access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Conductor New Track safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Conductor New Track use?

Conductor New Track is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Conductor New Track use?

About 3.6k tokens (SKILL.md is roughly 14k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Conductor New Track?

Skills that share tags, products or a category with Conductor New Track: Trellis Session Insight (mindfold-ai/Trellis, 15k stars), Native Data Fetching (CherryHQ/cherry-studio-app, 4k stars), Debugging Executions (n8n-io/n8n, 207k stars) and Aoti Debug (pytorch/pytorch, 104k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Conductor New Track?

gemini-cli-extensions (a GitHub organization) maintains it in gemini-cli-extensions/conductor, which has 3,757 GitHub stars. The repository holds 6 skills in this directory. The repository was last updated on September 1, 2026.

Source: gemini-cli-extensions/conductor on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.