Orlix Write To Harness

Orlix Write To Harness

Use this skill before writing any repository knowledge or agent-harness surface.

Rules

• Keep AGENTS.md a concise router into docs/index.md, typed knowledge pages, and structured reports.
• Follow docs/ontology.md and docs/AGENTS.md for every knowledge mutation.
• Keep durable architecture, ownership, capabilities, outcomes, stories, and tasks in typed pages under docs/objects/ and docs/concepts/.
• Organize work as epic -> story -> task, and keep each work page in its matching todo/, doing/, or done/ folder.
• Keep current task selection, commands, evidence, failures, and readiness state in structured reports under Build/AgentHarness/.
• Record each independently verified PR #228 recovery checkpoint in IMPLEMENT.md, as required by the accepted plan. Link to raw structured evidence instead of duplicating it. Do not create parallel journals or status archives.
• Update docs/log.md, regenerate docs/index.md, and repair every consumer when knowledge changes.
• Keep OrlixKit as the public SDK and OrlixOS as the running OS. Preserve upstream Linux ownership and the separation between private native implementation and guest distribution artifacts.
• Edit .rulesync/ as the durable source for generated rules and subagents. Generate only the intended features so hand-maintained skills remain intact.
• Preserve and identify user-owned generated-file deltas before regeneration. Reapply or incorporate their intent without loss and record the preserved delta and generated diff in the checkpoint.
• Treat command wrappers identically to their underlying commands in hooks and rules.

Verification

Run make docs-check, make agent-harness-check, a stale legacy-path scan, and the focused tests for any changed hook or workflow.