ospec

Converts structured specification documents (SDD format) into executable code generation prompts by parsing document structure, extracting requirements, and mapping them to code generation contexts. Uses document metadata and hierarchical sections to maintain semantic relationships between specifications and generated code artifacts, enabling AI coding assistants to generate code that directly implements documented requirements.

Parses specification documents (markdown, SDD format) into abstract syntax trees, extracting sections, requirements, constraints, and metadata. Maps document structure to semantic units that can be queried and referenced by code generation pipelines. Handles nested sections, requirement hierarchies, and cross-references to build a queryable specification model.

Transforms extracted specification requirements into optimized prompts for AI coding assistants by selecting relevant sections, formatting constraints, and building context windows that maximize code generation quality. Uses document structure to prioritize high-level requirements, acceptance criteria, and constraints in the prompt, reducing token waste and improving model focus.

Maintains mappings between specification sections and generated code artifacts, enabling developers to trace which code implements which requirements and which requirements are covered by which code. Supports querying code to find its source requirements and querying requirements to find implementing code, with metadata about coverage and implementation status.

Orchestrates multi-step workflows that combine specification parsing, prompt generation, code generation, and traceability tracking into automated pipelines. Manages state across workflow steps, handles errors, and coordinates between specification documents and AI coding assistants. Supports both synchronous generation and asynchronous workflows with callback handling.

Analyzes specifications to identify incomplete requirements, missing acceptance criteria, and coverage gaps. Validates specification structure against SDD standards and checks for consistency. Generates coverage reports showing which requirements have been addressed by generated code and which remain unimplemented.

Generates code across multiple files while maintaining specification context and consistency. Manages dependencies between generated files, ensures cross-file references are correct, and tracks which specification sections apply to which files. Handles file organization, naming conventions, and directory structure based on specification organization.

Tracks changes to specifications over time, maintains version history, and identifies what changed between specification versions. Enables developers to understand how specifications evolved and what code changes are needed when specifications are updated. Supports diffing specifications and generating change summaries.

Provides abstraction layer for integrating with multiple AI coding assistant APIs (OpenAI, Anthropic, etc.) and LLM providers. Handles API authentication, request formatting, response parsing, and error handling. Supports streaming responses, token counting, and cost tracking. Abstracts provider-specific differences to enable switching between providers.