# Harbor Self-Evolving

> Continuous evaluation and controlled self-evolution for DeepSeek Harness Agents.

---

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

---

Section pages:

- [Engineering notes](/harbor-self-evolving/blog/): Design notes about evidence, evaluation architecture and the boundaries behind Harbor Self-Evolving.
- [From real Sessions to controlled promotion](/harbor-self-evolving/book/): A six-chapter guide from real failure evidence to one comparable Candidate regression and an external deployment handoff.
- [Documentation](/harbor-self-evolving/docs/): Install Harbor Self-Evolving, understand the evidence model, run both evaluation paths and inspect exact contracts.
- [Product](/harbor-self-evolving/product/): Understand the three deliverables, two evaluation paths, design principles, history and current limits.
- [Releases and evidence](/harbor-self-evolving/releases/): Formal release identities, shipped capabilities, evidence types and explicit verification limits.