sdd

Orchestrates a multi-phase pipeline from Jira issue to spec, implementation, review, tests, and commit.

2|Updated Jun 5, 2018
One-click install
npx skills add https://github.com/Servoy/custom-rendered-components --skill sdd-servoy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sdd
Source: https://github.com/Servoy/custom-rendered-components/tree/main/.opencode/skills/sdd
Command: npx skills add https://github.com/Servoy/custom-rendered-components --skill sdd-servoy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Turning a Jira ticket into reviewed, tested, and committed code normally requires juggling investigation, spec writing, coding, code review, and test generation by hand. This Skill automates that entire Spec-Driven Development pipeline as an orchestrated sequence of isolated subagent phases with human approval gates. ## Core Features & Use Cases - Root-Cause Triage: An independent triage agent challenges the ticket's premise, investigates the codebase and git history, and issues a PROCEED, NO_ACTION, or NEEDS_INPUT verdict before any spec is written. - Context-Isolated Phases: PM Agent, Coding, Code Review, Test Generation, and Test Review each run as fresh subagents that only receive the information they need, preventing cross-phase bias. - Human Approval Gates: Every major transition (triage decision, spec approval, review fixes, final commit) pauses for explicit user confirmation, including an optional flow to post clarifying questions back to the Jira issue. - Use Case: Run the pipeline on a ticket like SVY-21080 to get a triage report, an approved spec under docs/, an implementation matching project conventions, a generated JUnit test suite, and a formatted commit. ## Quick Start Run the sdd pipeline on Jira issue SVY-21080 and walk me through each approval gate until the final commit.

Frequently Asked Questions about sdd

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a spec-driven development pipeline from a Jira ticket?▼

Invoke the pipeline with a Jira issue key such as SVY-21080, optionally followed by extra context. The orchestrator runs triage, spec writing, coding, code review, test generation, and test review phases, pausing at human approval gates between each step.

What does the triage phase do before writing a spec?▼

The triage agent investigates the codebase and git history to find the actual root cause, challenges the ticket's proposed solution, and issues a PROCEED, NO_ACTION, or NEEDS_INPUT verdict. A human always confirms the verdict before the pipeline continues.

Can the pipeline post clarifying questions back to Jira?▼

Yes. When triage returns NEEDS_INPUT, the orchestrator can compose the reporter-facing questions as a Jira comment, show the exact text for approval, and post it via the Jira REST API using curl with a configured auth token.

Why does each pipeline phase run as an isolated subagent?▼

Context isolation prevents bias between phases. The coder only sees the spec, the reviewer only sees the spec plus actual code, and the test generator never sees review findings, so each phase forms an independent judgement.

What happens when the code review or test review finds problems?▼

If a review returns CHANGES NEEDED, you can auto-fix by spawning a coding agent with the review findings, fix manually, or override. The review phase then re-runs and repeats until it returns APPROVED or you override.

When should I not use this automated development pipeline?▼

Avoid it for trivial one-line fixes where full triage, spec, and review overhead is unjustified, or when the Jira ticket lacks enough detail and the reporter cannot answer clarifying questions. The triage gate itself may also recommend no action.