DLG Labs logo

LEONA by DLG Labs

Public benchmark dashboard for controlled repair benchmarks, real OSS replay probes, and adaptive true-LLM repair attempts. This dashboard exposes summaries, pass/fail metrics, repair telemetry, retry counts, rollback events, diffs, repo traces, and the claim boundary between deterministic replay and true LLM repair. Codex-assisted operation exercised LEONA's repair infrastructure; LEONA generated the governed benchmark telemetry.

Benchmark Selection

Operator Disclosure

Codex was used as a development/operator assistant to launch commands, inspect outputs, and guide repository maintenance. LEONA performed the governed repair benchmark execution through its own repair pipeline, validation layer, rollback system, mutation constraints, and telemetry generation.

Retry Counts and Rollback Events

Category Coverage

Repo Traces and Case Results

Case Result Retries Rollback Duration Files Commit