Spec Auditing with Quality Guard - Axiooma

Spec Auditing with Quality Guard

The Quality Guard is a core component of the Axiooma platform. It operates as an automated requirements auditor that checks your specification documents (Specs) and rates them on a scale from 1.0 to 10.0. It flags logical loopholes, ambiguities, and missing edge cases before coding begins.

Evaluation Dimensions

Quality Guard analyzes your specifications across three critical engineering dimensions:

1. Information Completeness

• Objective: Ensure that the spec explains the what, the why, and the for whom of the proposed feature.
• Criteria: Checking for clear business goals, data validation rules, explicit error handling paths, and defined system boundaries.

2. Ambiguity & Inconsistency Detection

• Objective: Prevent developers from misinterpreting rules due to vague language or contradictory statements.
• Criteria: Use of clear imperative verbs, removal of subjective terms (e.g. “fast”, “sometimes”, “as needed”), and logical harmony across sections.

3. Solution Separation (Technical Isolation)

• Objective: Ensure the spec describes the business behavior and constraints, without dictating detailed UI layouts (e.g., colors, specific buttons) or strict database/API implementation choices.
• Criteria: The design and technical choices must be left to designers and developers; the requirements document must focus on logical behavior.

Quality Verdicts

After the audit finishes, one of the following localized verdicts is assigned based on the quality score:

• READY FOR DESIGN (Apto para Diseño - Score ≥ 9.0): The specification is robust, complete, and free of ambiguities. It is mature enough to be handed off to developers and designers.
• REVISION REQUIRED (Revisión Requerida - Score 5.0 - 8.99): The document is built on a good foundation but lacks key details (e.g., error paths or boundaries). The AI assistant compiles a clean Todo checklist to help you fix it.
• REJECTED (Rechazado - Score < 5.0): The specification is highly incomplete, blends business logic with premature implementation, or contains multiple logical errors. It requires a deep rewrite.

Improving Your Spec Score

1. Review the detailed feedback points generated by Quality Guard.
2. Open the AI Assistant chat panel on the right side of the editor.
3. Prompt the assistant with specific tasks, such as: “Help me write the missing data validation rules identified by Quality Guard”.
4. Apply the AI suggestions and run the quality audit again until you reach the desired score.