← Back to Skills Marketplace
905
Downloads
0
Stars
4
Active Installs
1
Versions
Install in OpenClaw
/install subagent-driven-development-2
Description
Execute implementation plans by dispatching a fresh subagent per task with two-stage review (spec compliance then code quality). Use when you have an implementation plan with mostly independent tasks and want high-quality, fast iteration within a single session.
Usage Guidance
This is an instruction-only orchestration skill that looks coherent for per-task development work, but it omits some concrete pieces and implicitly requires repository/git/test-run privileges. Before installing or using it:
- Confirm where the referenced prompt templates (references/implementer-prompt.md, spec-reviewer-prompt.md, code-quality-reviewer-prompt.md) come from — they are not included in the package. Ask the publisher for the exact prompt text.
- Only run this skill in an environment you trust. It expects the controller to run tests, commit code, and operate on a repo; granting those capabilities to an agent gives it the ability to modify source and push commits.
- If you plan autonomous invocation, limit the agent's repository/write/network permissions or test in a sandbox first.
- If you need stricter guarantees, request an explicit install artifact or prompts and a clear dispatch mechanism (how subagents are spawned), and consider adding guardrails (dry-run mode, approval steps, scoped commit access).
These gaps don't imply malicious intent, but they are practical inconsistencies you should clarify before handing the skill live repository or execution privileges.
Capability Analysis
Type: OpenClaw Skill
Name: subagent-driven-development-2
Version: 0.1.0
The skill bundle describes a workflow for an AI agent to orchestrate subagents for software development tasks. The `SKILL.md` file contains detailed instructions for the agent, including responsibilities, quality gates, and guardrails (e.g., 'NEVER Do'). There is no evidence of prompt injection attempts to manipulate the agent into performing unauthorized actions, exfiltrating data, or executing malicious commands. The instructions are entirely aligned with the stated purpose of managing a development process.
Capability Assessment
Purpose & Capability
Name/description (dispatch fresh subagents, two-stage review) align with the SKILL.md content. The skill does not request unrelated credentials, binaries, or config paths.
Instruction Scope
The runtime instructions assume the controller/agent will implement dispatching subagents, run tests, make commits, and reference git SHAs — but the skill does not include the referenced prompt files (references/*.md) and does not declare or limit how subagents are dispatched. The instructions therefore implicitly assume repository, filesystem, and git/tool access that are not made explicit. This is coherent for a developer orchestration skill but incomplete and gives the agent broad discretion.
Install Mechanism
No install spec and no code files — lowest-risk installation model. The README shows an npx install example, but no actual install script or package is provided in the registry entry; that's a documentation/incompleteness issue rather than an install risk.
Credentials
The skill declares no required environment variables, credentials, or config paths. However, the instructions imply the need for access to the working repository, git, test runner, and possibly external issue trackers or CI; those are not requested explicitly.
Persistence & Privilege
always is false and model invocation is allowed (platform default). The skill does not request persistent presence or attempt to modify other skills or global agent settings in the provided text.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install subagent-driven-development-2 - After installation, invoke the skill by name or use
/subagent-driven-development-2 - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.0
Initial public release of subagent-driven-development skill.
- Dispatches a fresh subagent for each task in an implementation plan, enforcing two-stage review: spec compliance then code quality.
- Optimized for fast, high-quality iteration on plans with mostly independent tasks, all within a single session.
- Provides a detailed workflow and controller responsibilities to avoid context pollution and ensure review order.
- Includes prompt templates for implementer, spec reviewer, and code reviewer subagent roles.
- Outlines when (and when not) to use the skill, common pitfalls, and strict quality/review guidelines.
Metadata
Frequently Asked Questions
What is Subagent Driven Development?
Execute implementation plans by dispatching a fresh subagent per task with two-stage review (spec compliance then code quality). Use when you have an implementation plan with mostly independent tasks and want high-quality, fast iteration within a single session. It is an AI Agent Skill for Claude Code / OpenClaw, with 905 downloads so far.
How do I install Subagent Driven Development?
Run "/install subagent-driven-development-2" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Subagent Driven Development free?
Yes, Subagent Driven Development is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Subagent Driven Development support?
Subagent Driven Development is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Subagent Driven Development?
It is built and maintained by wpank (@wpank); the current version is v0.1.0.
More Skills