harness-work

Autonomous feature work automates end-to-end implementation by following a structured orient-stabilize-select-implement-self-check-verify protocol, ensuring traceable progress and test validation before committing. It selects a single highest-priority feature and updates progress docs throughout the cycle.

Structured progress tracking for autonomous coding follows an orient-stabilize-select-implement-self-check-verify protocol. It enforces disciplined, incremental feature work by running tests, fixing blockers, updating progress docs, and delegating verification to ensure reproducible commits.

Yes, this autonomous feature work protocol is designed exclusively for harness-engineered repositories. It requires a managed environment with a defined command log, progress documentation, and recent commit history to orient the agent and enforce structured feature implementation.

Test and build validation is enforced during the stabilize and self-check phases. The agent runs tests, fixes blockers before implementing new work, performs self-checks post-implementation, and delegates final verification to an external verifier to ensure quality and consistency before committing.

Managing incremental feature work across sessions requires a structured protocol that orients the agent to recent commits and progress docs, stabilizes the environment, and implements one feature per cycle. This maintains traceable, reproducible progress by enforcing test validation and progress tracking.

Avoid using an autonomous feature workflow outside harness-engineered repositories or for tasks not fitting a single-feature cycle. The strict orient-stabilize-select-implement-self-check-verify protocol enforces structured progress tracking and test validation, making it unsuitable for unstructured or multi-feature coding tasks.