technical-specification · aj-geddes — Claude Code skill: specs and requirements
Gives your agent a structure for technical specs, PRDs, ADRs and RFCs: requirements, architecture, acceptance criteria. Useful before you start coding a feature.
npx skills add aj-geddes/useful-ai-prompts --skill technical-specification -a claude-codeWhat technical-specification does
The skill lays out a document skeleton: status and version, an executive summary with problem and solution, a background section, then functional and non-functional requirements, database schema, API data models, authentication endpoints and rate limiting. In the repo these live as separate files under references/, with SKILL.md as the entry point.
A markdown template ships with it, and the agent fills it in for your task. There is also a phased plan: core authentication first, then email verification, social login, security work. A do-not list covers the usual gaps — vague requirements, skipped non-functional items, no testing strategy, no monitoring.
The skill does not read your code or write it. It works on the document text, so the output is only as good as the problem description you feed it. If the requirements are still fuzzy in your head, you will get headings and little else.
What it helps with
- describe a feature before development starts
- draft a PRD with acceptance criteria
- record an architecture decision as an ADR
- design a database schema and API models
- break work into phases with risks and mitigations
Example prompts for the agent
- Write a technical spec for email signup: requirements, DB schema, endpoints
- Draft an ADR for our message queue choice, with alternatives and consequences
- Put together an RFC for a new public API with rate limiting
How to install technical-specification
1Skills CLI (Vercel)
npx skills add aj-geddes/useful-ai-prompts --skill technical-specification -a claude-code2Manually: put the files in place
~/.claude/skills/technical-specification/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/technical-specification/
1Skills CLI (Vercel)
npx skills add aj-geddes/useful-ai-prompts --skill technical-specification -a codex2Manually: put the files in place
.agents/skills/technical-specification/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add aj-geddes/useful-ai-prompts --skill technical-specification -a cursor2Manually: put the files in place
.cursor/skills/technical-specification/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install aj-geddes/useful-ai-prompts technical-specification2Manually: put the files in place
.github/skills/technical-specification/SKILL.md1Skills CLI (Vercel)
npx skills add aj-geddes/useful-ai-prompts --skill technical-specification -a gemini-cli2Platform CLI
gemini skills install https://github.com/aj-geddes/useful-ai-prompts --path skills/technical-specificationManually: put the files in place
.devin/skills/technical-specification/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add aj-geddes/useful-ai-prompts --skill technical-specification -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about technical-specification
How do I install technical-specification in Claude?
Run npx skills add aj-geddes/useful-ai-prompts --skill technical-specification -a claude-code. The skill file lands at ~/.claude/skills/technical-specification/SKILL.md.
What does technical-specification do?
It helps the agent produce a technical spec: requirements, architecture, data schema, API contracts and acceptance criteria. Works for PRDs, ADRs and RFCs too.
Which platforms are supported?
Besides Claude there are installs for Codex, Cursor, GitHub Copilot, Gemini CLI and OpenCode. GitHub Copilot uses a different command: gh skill install aj-geddes/useful-ai-prompts technical-specification.
Source text
SKILL.md · MIT
--- name: technical-specification description: > Create detailed technical specifications, requirements documents, design documents, and system architecture specs. Use when writing technical specs, requirements docs, or design documents. --- # Technical Specification ## Table of Contents - [Overview](#overview) - [When to Use](#when-to-use) - [Quick Start](#quick-start) - [Reference Guides](#reference-guides) - [Best Practices](#best-practices) ## Overview Create comprehensive technical specifications that define system requirements, architecture, implementation details, and acceptance criteria for software projects. ## When to Use - Feature specifications - System design documents - Requirements documentation (PRD) - Architecture decision records (ADR) - Technical proposals - RFC (Request for Comments) - API design specs - Database schema designs ## Quick Start Minimal working example: ```markdown # Technical Specification: [Feature Name] **Document Status:** Draft | Review | Approved | Implemented **Version:** 1.0 **Author:** John Doe **Date:** 2025-01-15 **Reviewers:** Jane Smith, Bob Johnson **Last Updated:** 2025-01-15 ## Executive Summary Brief 2-3 sentence overview of what this spec covers and why it's being built. **Problem:** What problem are we solving? **Solution:** High-level description of the solution **Impact:** Expected business/user impact --- ## 1. Background ### Context Provide background on why this feature is needed: // ... (see reference guides for full implementation) ``` ## Reference Guides Detailed implementations in the `references/` directory: | Guide | Contents | |---|---| | [Functional Requirements](references/functional-requirements.md) | Functional Requirements | | [Non-Functional Requirements](references/non-functional-requirements.md) | Non-Functional Requirements | | [Database Schema](references/database-schema.md) | Database Schema | | [API Data Models](references/api-data-models.md) | API Data Models | | [Authentication Endpoints](references/authentication-endpoints.md) | Authentication Endpoints | | [Rate Limiting](references/rate-limiting.md) | Rate Limiting | | [Phase 1: Core Authentication](references/phase-1-core-authentication.md) | Phase 1: Core Authentication (Week 1-2), Phase 2: Email Verification (Week 3), Phase 3: Social Login (Week 4), Phase 4: Security Features (Week 5) (+1 more) | ## Best Practices ### ✅ DO - Include acceptance criteria for each requirement - Provide architecture diagrams - Document API contracts - Specify performance requirements - List risks and mitigations - Include implementation timeline - Add success metrics - Document security considerations - Version your specs - Get stakeholder review ### ❌ DON'T - Be vague about requirements - Skip non-functional requirements - Forget about security - Ignore alternatives - Skip testing strategy - Forget monitoring/observability - Leave questions unanswered
© aj-geddes · MIT · aj-geddes/useful-ai-prompts
Similar skills
Everything for Writing and docs →technical-specification by other authors comes first.