- Home
- Skills
- Catlog22
- Claude Code Workflow
- Spec Generator
spec-generator_skill
- TypeScript
1.3k
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill catlog22/claude-code-workflow --skill spec-generator- README.md3.0 KB
- SKILL.md12.2 KB
Overview
This skill generates a complete, traceable specification package through a six-phase document chain that produces a product brief, PRD, architecture docs, epics, and a readiness report. It orchestrates multiple CLI analysis agents in parallel, supports interactive confirmation points, and creates a structured output directory ready for handoff to execution workflows. The workflow is resumable and template-driven, focused exclusively on documentation generation.
How this skill works
The generator parses input (text or file), creates a session and seed analysis, then runs six sequential phases: discovery, product brief, requirements (PRD), architecture, epics & stories, and readiness check. Multiple CLI tools provide product, technical and user perspectives in parallel; outputs are synthesized, validated, and written to templated markdown files with YAML frontmatter. A spec-config state file tracks progress and enables resume and auto modes; final outputs are packaged for downstream execution workflows.
When to use it
- You need a complete, traceable product specification from idea to executable epics.
- Preparing handoff artifacts for planning workflows, issue creation, or feature delivery.
- Running a repeatable, template-driven spec process with traceable decisions and ADRs.
- When multiple perspectives (product, technical, user) must be synthesized into one spec package.
Best practices
- Read the required document and template standards before starting to ensure correct formatting and quality gates.
- Start interactive mode for checkpoints; use the auto (-y) flag only when defaults are acceptable.
- Use the continue (-c) flag to resume long-running or interrupted sessions from the last checkpoint.
- Keep the workspace under the session directory to preserve state and traceability for audits.
- Use the multi-CLI perspectives to resolve conflicts early and capture acceptance criteria and non-functional requirements explicitly.
Example use cases
- Generate a product brief and PRD for a new feature idea and export epics for sprint planning.
- Create architecture decisions and ADRs for a proposed tech redesign, including component and stack summaries.
- Produce a readiness report and one-page executive summary for stakeholder review before development kickoff.
- Decompose PRD requirements into epics and stories with Mermaid dependency maps for roadmap planning.
FAQ
No. It only generates documentation and handoff artifacts. Execution is performed by separate workflow tools after handoff.
How do I resume a partially completed spec run?
Use the continue flag to read the session state file; the system resumes from the first incomplete phase recorded in spec-config.json.
What happens if a CLI call fails during a phase?
There are fallbacks: remaining CLI tools, local analysis, or degraded mode. The pipeline continues where possible and generates partial outputs if necessary.