blueprint

Orchestrates spec-driven feature development through subagent implementation, verification, and simplification phases.

Updated May 18, 2024
One-click install
npx skills add https://github.com/joshhornby/dotfiles --skill blueprint-joshhornby
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: blueprint
Source: https://github.com/joshhornby/dotfiles/tree/main/.claude/skills/blueprint
Command: npx skills add https://github.com/joshhornby/dotfiles --skill blueprint-joshhornby

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Building features without a clear spec leads to wasted iterations, scope creep, and unverified code. This Skill enforces a structured spec-first workflow where every implementation is tested, reviewed, and simplified before completion. ## Core Features & Use Cases - Spec-First Planning: Collaborates with the user to write a detailed spec.md covering success criteria, scope, architecture decisions, tasks, and test plans before any code is written. - Parallel Subagent Execution: Spawns implementation, test-running, code-review, and lint-checking subagents that work concurrently with explicit pass/fail output contracts. - Verification Gates: Requires tests, review, and static analysis to all pass before proceeding, with fix subagents and bounded retry loops when issues are found. - Use Case: Ask the agent to build a new API endpoint from scratch; it drafts a spec for your approval, implements tasks via subagents, verifies against the spec, then simplifies the code while re-running tests. ## Quick Start Ask the agent to build your feature using the blueprint workflow, starting with a spec for your approval.

Frequently Asked Questions about blueprint

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

FAQPage Schema
How do I build a feature with a spec-first AI workflow?

Start by collaborating on a spec.md covering problem statement, success criteria, scope, architecture, tasks, and test plan. After your explicit approval, subagents implement each task, then verification subagents run tests, review code, and check types before simplification.

How does the blueprint skill use subagents for implementation?

Each implementation task gets a subagent prompt with the task description, relevant spec context, target files, and success criteria. Independent tasks run in parallel, dependent tasks run sequentially, and each subagent reports TASK_COMPLETE or BLOCKED with a reason.

What happens when verification fails during development?

If tests, code review, or lint checks fail, a fix subagent is spawned with the specific issues, then the failed verification re-runs. The loop repeats up to five iterations, after which remaining issues are reported instead of blocking indefinitely.

When should I not use a spec-driven multi-phase workflow?

Avoid it for trivial one-file changes, quick bug fixes, or exploratory prototypes where a full spec and verification gates add overhead without value. It fits multi-phase features where correctness and review matter more than speed.

Does the simplification phase change code behavior?

No. Simplification subagents remove dead code, over-abstraction, and noisy comments while preserving all existing behavior and tests. After simplification, the full verification phase re-runs to confirm nothing broke.