# Workflows

> Diagnose real Sessions, run Candidate regressions, and govern Evaluators.

---

LLMS index: [llms.txt](/harbor-self-evolving/llms.txt)

---

- [Historical diagnosis](https://istarwyh.github.io/harbor-self-evolving/docs/workflows/historical/)
- [Candidate evaluation and promotion](https://istarwyh.github.io/harbor-self-evolving/docs/workflows/candidate/)
- [Evaluator governance and meta-evaluation](https://istarwyh.github.io/harbor-self-evolving/docs/workflows/evaluator/)
{.cards}

---

Section pages:

- [Evaluate recent Sessions](/harbor-self-evolving/docs/workflows/historical/): Turn bounded, redacted recent DSH Sessions into a non-promotion Historical evaluation Job.
- [Candidate evaluation and promotion](/harbor-self-evolving/docs/workflows/candidate/): Freeze identities, run a comparable regression, and obtain a deterministic promotion recommendation.
- [Evaluator governance and meta-evaluation](/harbor-self-evolving/docs/workflows/evaluator/): Inspect, version and independently test the Evaluator instead of treating the scoring rule as an oracle.