Heuristic
Turn repeated human corrections into guardrails with wider reach
If humans repeat the same comment every week, the feedback is trapped at the wrong layer.
When it fits
- Reviewers keep making the same correction to AI-generated work.
When to avoid it
- Do not institutionalize a reviewer preference until the team agrees it is a real requirement.
Why it matters
Classify the correction. If it is deterministic, encode it as a lint rule, test, schema or script. If it is a stable convention, put it in discoverable project documentation or a skill. If it needs judgment, create a focused review rubric or reviewer agent and keep human escalation. Aim to move recurring feedback from one artifact to the system that produces many artifacts.
An example
Instead of repeatedly telling an agent not to modify generated files, enforce the rule in repository instructions and CI.
Check your result
The next similar task receives the correction before or during generation, not only after a human spots it again.
Keep this limit in mind
- Do not institutionalize a reviewer preference until the team agrees it is a real requirement.
Connected ideas
Useful withTurn a verified AI failure into a regression test
Evidence and sources
AI Engineer harness guidance recommends turning repeated human feedback into durable documentation, lint rules, structural tests or specialized review agents.
Automated guardrails can encode a bad preference; keep ownership and review for consequential rules.
Harness Engineering: How to Build Software When Humans Steer, Agents Execute · Make quality legible and enforceable