Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Behavioral guardrails for Cavekit agents. Four principles — think before coding, simplicity first, surgical changes, goal-driven execution — that prevent over-engineering, silent assumptions, scope creep, and unfocused work. Every task-builder, reviewer, planner, and inspector must internalize these before writing a single line. Trigger phrases: "guardrails", "karpathy", "scope creep", "over-engineering", "stop adding features", "surgical fix".
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 40% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 95% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 22% | 0% |
Four rules. Load them into context at the start of every task. The reviewer enforces them as a Pass-1 filter before it looks at code quality.
Before the first edit, write down:
unverified, flag NEEDS_CONTEXT and ask — do not guess.
test, check, or observable behaviour. If a criterion is not verifiable, propose a sharpening via the revision skill (automated-trace subsection), not a vague attempt.
Refusing to produce code is allowed. A task with unknown scope is a spec bug, not a coding task.
The correct amount of code is the minimum that meets the acceptance criteria.
abstraction with two configuration knobs.
If the diff is larger than the acceptance criteria seem to demand, explain why in the commit body. If you cannot, trim the diff.
Every line in the diff must trace back to an acceptance criterion. Touching code outside the task's owned files is justified only when a requirement forces it. Examples of violations:
If you see a real bug in adjacent code, log it to .cavekit/history/backprop-log.md as a candidate kit item and keep it out of this task's diff.
Transform vague tasks into verifiable success criteria before execution.
exact files to inspect. "Make sure it works" is not a plan.
each AC, the verification step, and the observed result.
any of: undeclared assumptions, diff lines unjustified by an AC, out-of-scope edits, or unreachable verification steps.
bugs and block planning.
or hand-waved.
You are probably over-confident about a shortcut that will cost more than the delay of asking. Stop and note the tension in the commit body or in the build log so the reviewer can judge.
Other measured skills in the registry, with their headline benchmark lift.