ada-srs-to-detailed-design

Generate detailed design chapters from layered SRS documents with requirement traceability.

Updated Jul 23, 2026
One-click install
npx skills add https://github.com/wubing7755/Ada --skill ada-srs-to-detailed-design-wubing7755
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ada-srs-to-detailed-design
Source: https://github.com/wubing7755/Ada/tree/main/skills/software-development/ada-srs-to-detailed-design
Command: npx skills add https://github.com/wubing7755/Ada --skill ada-srs-to-detailed-design-wubing7755

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Writing a software detailed design (SDD) chapter from a layered Software Requirements Specification is error-prone: entries get derived from the wrong requirement layer, descriptions merely restate requirements instead of explaining how to implement them, and coverage gaps or numbering errors slip through review. This Skill enforces a disciplined derivation workflow so every design entry traces to an implementation requirement and passes automated validation. ## Core Features & Use Cases - Layered derivation rules: Reads all three SRS layers (user/development/implementation requirements) but derives design entries only from implementation requirements, flagging uncovered performance/capacity requirements as coverage risks. - Phased delivery gates: Produces a 14-field requirement analysis list and a 7-field requirement-to-design mapping table for user confirmation before writing formal four-field design entries (ID, name, description, source). - Automated validation scripts: Ships Python scripts that check ID continuity, four-field completeness, source-requirement legality, full implementation-requirement coverage, constraint-copying similarity, and format compliance. - Use Case: Given a GJB 438B-style layered SRS for a multi-tool resource editor, generate the "5 Detailed Design" chapter with 60+ self-contained SD entries, a mapping table proving 91/91 implementation-requirement coverage, and pass all validation scripts before delivery. ## Quick Start Generate the detailed design chapter from my layered SRS file, deriving design entries only from implementation requirements and validating coverage with the included scripts.

Frequently Asked Questions about ada-srs-to-detailed-design

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

FAQPage Schema
How do I generate a detailed design document from an SRS?

Read all three SRS layers, then derive design entries only from implementation requirements, organized by module rather than requirement order. Run a requirement-analysis phase first (14-field list), confirm a mapping table with the user, then write four-field entries and validate with the included scripts.

How to keep design descriptions from just restating requirements?

Design descriptions must answer HOW, not WHAT: include trigger, processing steps, data flow, mechanism collaboration with SD references, result state, and exception boundaries. The test is whether deleting the description would leave the implementer unable to code the entry.

Can design entries source from user or development requirements?

By default no; only implementation requirements appear in the source field. The exception is an explicit user decision to derive performance or capacity development requirements directly, which must be recorded in the mapping table and noted as a coverage exception.

What validation checks run before delivering the design document?

Run validate_sdd.py to check ID continuity, four-field completeness, source IDs belonging to implementation requirements, and full coverage. For four-field-only deliveries also run check_design_format.py, and use check_constraint_copy.py to detect constraints copied verbatim from SRS text.

Why does requirement coverage checking report false missing items?

SRS Source lines like "REQ-F-062 (parent name)" are parent-requirement references, not implementation requirements; parsers must anchor to whole-line ID matches. Module boundaries are H2 headings, so stopping collection at H3 level causes cross-module miscounts.