Rosetta Test Script Engine

7 sections

0%
Section 1 Walking Skeleton Section 2 Record / Write API (SSOT) Section 3 Verification Coverage Section 4 Measurement + Benchmark Engine + Fairness Section 5 Pipeline Orchestration + Retry Cure Section 6 Re-author rosetta-stress-test Section 7 Thin-Skill Rewrite of /rosetta-test
Overview

Migrate the /rosetta-test skill into a fully script-driven engine: scripts own the ENTIRE mechanical rosetta pipeline deterministically (build, run, benchmark, measure deltas, 3-way checksum, dual-exec / leak / ARC verification, record). The LLM is used ONLY to author the idiomatic .ori program + microbench + tests and to analyze (ergonomics scoring, bug RCA, decisions, fixes).

Architectural migration with no approved proposal (script-first.md / workflow-ownership.md applied to /rosetta-test). Realizes the §AIMS + §Compiler competitive-benchmark missions via a deterministic, reproducible measurement harness. Strict hard-gate perf posture: a loss against C++ -O3 (compiled) or Python3 (interpreted) is a compiler/AIMS/interpreter defect routed through /add-bug + /fix-bug, never an accepted result.

○

Planned

7 sections
Section 1 Not Started

Walking Skeleton

0/2 tasks
Section 2 Not Started

Record / Write API (SSOT)

0/3 tasks
Section 3 Not Started

Verification Coverage

0/2 tasks
Section 4 Not Started

Measurement + Benchmark Engine + Fairness

0/5 tasks
Section 5 Not Started

Pipeline Orchestration + Retry Cure

0/3 tasks
Section 6 Not Started

Re-author rosetta-stress-test

0/3 tasks
Section 7 Not Started

Thin-Skill Rewrite of /rosetta-test

0/3 tasks