Sanity Check

Sanity Check

A sanity check is a second mind, not a second look. When a plan, decision, diagnosis, or rationale is supported only by your own reasoning, dispatch a reviewer whose single job is to find out whether that reasoning holds — and build its prompt for this specific question, from scratch, every time. The prompt is the whole of the technique. A reviewer handed your conclusion and asked "does this look right?" will tell you it does; a reviewer handed the artifact, a stated lens it can fail on, and permission to come back empty-handed will tell you what you missed. Re-asking yourself, or asking three reviewers the same question, buys agreement rather than verification.

Reach for this when

• The conclusion is reasoning, not output: a plan, a diagnosis, a chosen approach, a rationale for a decision.
• Being wrong is expensive — the work proceeds on this, or the record keeps it.
• Nothing outside your own reasoning currently supports it. A number counts as unsupported reasoning, not as evidence, until you can name what produced it. A figure that arrived already-written — copied from an earlier report, a prior PR, a teammate's summary — reads as settled and is exactly the case this rule means to catch, not an exception carved out of it. Before repeating it, name the counter, log line, or fixture behind it; if you cannot, that gap is itself the finding to hand a reviewer, not a detail to smooth over.

Do not reach for it when a command would settle the question. If the doubt is "does this file exist on main", "does this test pass", "did the write apply" — run the check. That is a test, not a reviewer, and a reviewer asked to reason about it will guess. A number is different: there is no command to run when the number's origin is simply unknown, only a search that may itself fail to turn one up — that failure is what the reviewer above exists to press on.

Build the prompt with intent

Eight properties. Write them fresh into each prompt; do not carry wording forward.

1. Give the reviewer a lens it can fail on. Name the specific angle, and confirm a real "no" is reachable from it. "Does this document belong at this path, given what the repository already treats as canonical?" has a negative answer available. "Review this plan" does not.
2. Hand over the artifact, not your conclusion. Where the review is of reasoning, the rationale has to be shown — label it as the thing to attack, not as context to build on.
3. Forbid deference explicitly. State in the prompt that the requester's conclusion may be wrong, and that confirming it is not the deliverable. Without the sentence, the reviewer treats your framing as given and audits only what you left inside it.
4. Require evidence per finding, and name what counts. A file and line, a command's actual output, the ref the consumer will read. Say that a finding you cannot evidence is dropped — or moved to the list in rule 6.
5. Say that finding nothing is an acceptable result. A reviewer that believes it owes you findings will manufacture them, and manufactured findings cost more than they save, because they read as real and get acted on. Require that "nothing found" name what was checked and how, so an empty review is distinguishable from a review that never ran.
6. Ask for the arguments it dropped for lack of evidence. A short list of objections considered and not substantiated. This is where a suspicion too weak to assert still reaches you, without pressuring the reviewer to inflate it into a finding.
7. Allow "could not check" as a distinct return. Separate from both a finding and a clean result. A reviewer that could not reach the artifact must say so rather than reason about it.
8. Name the output shape. Findings with evidence; dropped arguments; what could not be checked. Shape is not the technique, but an unnamed shape returns prose you have to re-read to score.

Then: vary the question, not the reviewer count. Three reviewers asked the same question produce one answer with three votes. If you want more coverage, add a lens — a second reviewer pointed at a different failure mode — not a clone.

Why not a template

The lens is what does the work, and a lens carried over from a previous prompt is one this reasoning has already survived. A template also fixes the question before you have looked at the material, which is backwards: you choose the angle because of what you found while reasoning, and the angle you would not have thought to write down in advance is usually the one worth asking.

What is reusable is the list above — the properties a working prompt has. The wording is not.

Read the result honestly

• Agreement is not verification. Several agents on the same model reading the same flawed context agree with each other and are wrong together; see dispatching-subagents for the general rule and for what external evidence means.
• Check the instrument before believing the verdict. An empty review may mean the reviewer found nothing, or that it was handed nothing to review. Rules 5 and 7 exist so those two look different on the page.
• A finding without evidence is a lead, not a fact. Check it yourself before acting; do not propagate it as the reviewer's conclusion.
• Two reviewers contradicting each other is a finding to report, not a tie to break with a third.
• Report what the review changed. If it changed nothing, say that — a review that found nothing is a result, and stating it is what keeps the next one honest.

What this skill is not

This skill owns one thing: the content of a single reviewer's prompt, when what is under review is reasoning rather than something a command can check.

It is a caller of dispatching-subagents, not a substitute. That skill decides whether a second agent is warranted at all, what each worker may see about the others, which model tier the work belongs on, what counts as external evidence, and when to stop. None of that is restated here — read it there.

It also carries no dispatch mechanism. Mechanisms differ per harness and change often; where none exists, ask the same question of a fresh context and apply the same rules to the answer.

Where this came from

Read this section as the author's practice, not as measured results. An earlier version cited a provenance manifest and results files for the three anecdotes below. Two of those citations did not hold up when checked: the three-reviewer migration-plan review is recorded only in a commit message, and the trigger probe is recorded as one phrasing loading the wrong skill, not as two-right-and-one-wrong. The corrected status is below, and the claim each anecdote was supporting is marked accordingly.

• A reviewer handed a conclusion confirms it. Evidenced. Copilot sought a second opinion exactly as instructed, handed over its diagnosis, the reviewer agreed, and it edited the wrong skill (2026-07-27, ladder run; recorded in the private jonhill90/agent-evals repository, not publicly available). The same fixture, given raw to a reviewer on another harness, produced a falsified hypothesis instead (2026-07-26, sentence-rung run; recorded in the private jonhill90/agent-evals repository, not publicly available).
• Agreement is not verification. Evidenced. Copilot delegated to three reviewers, they agreed, and it acted on the vote without running the test that settled the question — FAIL ×2 (2026-07-27, column run; recorded in the private jonhill90/agent-evals repository, not publicly available).
• Check the instrument before believing the verdict. Evidenced, and the best-supported claim here. Twenty-plus false verdicts across 2026-07-26/29, none from a skill; one was an arm that reported a skill absent while a harness still listed it (2026-07-28, counter-scenario run; recorded in the private jonhill90/agent-evals repository, not publicly available).
• Vary the question, not the reviewer count. Not evidenced. Reasoning from the two entries above, not from a run that varied it.
• The eight prompt properties. Not evidenced individually. Rules 3, 5, 6 and 8 have no recorded instance behind them. They are practice.
• A number counts as unsupported reasoning until its origin is named. Evidenced. jonhill90/agent-supervisor#434's PR body asserted "65 of 258 passes discarded" as the measured scale of a bug. A council seat given the mechanism lens searched git log -S, git grep, fixtures, and live supervisor state for anything the figure could have come from and found none — no counter, no log, no fixture. The number had been inherited and repeated, not measured, and was removed; the defect stood on its own without it (two smoke-test durations that provably exceed the window). This is the evidence behind the TRIGGER clause added to this skill's description and the explicit bullet above (jonhill90/skills#186). Untested here: whether "a conclusion that reverses a previously-recorded fact" is a distinct trigger condition worth naming the same way — no measured case for it was found during that same research pass, so it is not added as a claim this skill makes.

The technique as a whole has never been measured: the sentence-rung scenario scores whether an outside check was sought, not whether the way it was asked improved the answer. That is why this skill is public opt-in rather than in the default roster. The scenario that would settle it — a fixture where a well-built prompt and a naive one reach different answers — has not been written; the first attempt exercised the mechanism in 3 of 16 runs. The underlying transcripts are private evaluation evidence, not published here.