doc-tspec-autopilot_skill

Overview

This skill automates generation and review of TSPEC test specifications from SPEC documents, producing UTEST, ITEST, STEST, and FTEST artifacts with TASKS-Ready scoring. It detects input type (SPEC or TSPEC), generates missing TSPEC suites in a nested folder layout, validates TASKS-Ready compliance, and runs an iterative review-and-fix cycle until quality targets are met. The pipeline supports both generate and review modes and emits structured markdown reports and traceability updates.

How this skill works

The autopilot parses the input identifier (SPEC-NN or TSPEC-NN) to decide whether to generate a TSPEC or run a review. For generation it extracts methods, interfaces, and data models from SPEC, plans coverage, and creates four test-type files per TSPEC (UTEST/ITEST/STEST/FTEST) with traceability tags and element IDs. It then validates TASKS-Ready scoring, applies auto-fixes where possible, and runs a review/fix loop using the doc-tspec-reviewer and doc-tspec-fixer skills until the target score or iteration limit is reached.

When to use it

• You have SPEC documents and need complete TSPEC suites with test types and traceability.
• You want to validate existing TSPEC directories and produce a TASKS-Ready quality report.
• You need automated conversion of SPEC coverage into UTEST/ITEST/STEST/FTEST artifacts.
• You require an iterative review-and-fix workflow to reach a TASKS-Ready threshold.
• You need standardized element IDs and cumulative traceability tags across test specs.

Best practices

• Provide SPEC identifiers in the format SPEC-NN so the autopilot can detect upstream inputs.
• Keep SPEC documents well-structured (methods, interfaces, data models) to improve extraction accuracy.
• Run validation with strict TASKS-Ready thresholds enabled when preparing for downstream TASKS generation.
• Maintain the nested TSPEC folder convention (TSPEC-NN_{slug}/) so reports and traceability updates land correctly.
• Limit parallel execution to the configured max_parallel to avoid resource contention.

Example use cases

• Generate TSPEC-05 from SPEC-05 when no TSPEC folder exists, producing four test-type files and a coverage matrix.
• Review TSPEC-01 to get a TASKS-Ready score breakdown and an R_review_report_v001.md with actionable fixes.
• Run an automated fix cycle: detect legacy element IDs, convert them to TSPEC.NN.TT.SS, regenerate the coverage matrix, and revalidate.
• Batch-process multiple SPEC inputs (SPEC-01,SPEC-02) to create or review corresponding TSPEC suites with iterative fixes.
• Produce a final TASKS-Ready report and update the project traceability matrix before TASKS generation.