Install
openclaw skills install @deciqai/cpa-return-reviewActivate when: a prepared 1040 or 1120-S is going to review; you are the second set of eyes on someone else's return; a firm wants one consistent review standard across reviewers; onboarding a preparer to the review bar; something feels off about a return and you want a systematic pass; user says 'review this return', 'before we file', 'second review', 'preparer checklist', 'did we miss anything on this return?'. Do NOT activate when: the return is not yet prepared; the question is a single technical treatment; the work is chasing missing documents rather than reviewing; the question is what an examiner would attack or whether a position invites audit rather than whether the return is right (use tax-prep-pre-file-audit-premortem instead — it applies to 1120-S as much as 1040). More: deciqai.com/c/cpa-return-review
openclaw skills install @deciqai/cpa-return-reviewNot legal or tax advice. No thresholds, limits, rates, or dates appear here — all of them expire. Verify every amount and every rule against current IRS/state guidance and current-year software logic. The practitioner who signs owns the return.
Activate when: a return is prepared and going to review; you are reviewing someone else's work; you want the same discipline applied by every reviewer in the firm. Do NOT activate when: you are preparing rather than reviewing; the question is one isolated treatment; the question is what an examiner would attack rather than whether the return is right — that is [tax-prep-pre-file-audit-premortem], on 1120-S as much as on 1040.
Errors that reach clients are rarely exotic. They are a carryover nobody re-checked, a state that quietly changed, a name spelled two ways, an election made last year and forgotten this year, and a number that is technically correct but obviously wrong next to last year's.
Reading a return top to bottom hunting for mistakes finds what you already suspect and nothing else. Different error classes hide from different kinds of attention, so the review is five separate passes with five different questions — not one careful read.
Run in order. Each sweep is its own pass.
Boring, and the source of the most embarrassing errors.
1120-S annual check: the S election is documented in the permanent file and nothing this year threatens it — an ineligible shareholder, a second class of stock created by a side agreement, a shareholder-count change. Checked once a year, and this is where.
Open last year beside this one. Every material change needs a one-sentence explanation.
Gate: a disappearing line item outranks a wrong number. Wrong numbers get caught by review; missing items get caught by nobody, because there is nothing on the page to look at.
For each material number, point at the document. If you cannot, that is the finding.
Gate: if a number's only source is "the software brought it forward," it has not been verified this year. Carryforward is a convenience, not evidence.
Look at the whole return as a stranger would.
That last question is the sweep compressed to one line. Any "no" stops delivery until resolved.
Write findings down. A silent fix teaches the preparer nothing and leaves no record.
Finding · Location (form/schedule/line) · Severity · Disposition · Resolved by/when
Severity: BLOCKER (cannot file) · CORRECTION (must fix) · QUERY (need an answer) · NOTE (next year).
Gate: nothing filed with an open BLOCKER or CORRECTION. A QUERY either resolves or converts to a documented acceptance with a stated rationale.
NOTE items are next year's planning and chase list, free to collect now while the context is in your head.
1120-S, second year, prepared by an experienced preparer. Sweeps 1, 3 and 4 came back clean.
Sweep 2 found a repairs-and-maintenance line at roughly six times prior year with no explanation. Not wrong on its face — the client had bought a building. Sweep 5 then asked the stranger question: would you explain that number out loud? The work turned out to include a roof replacement, a capital item sitting in a deduction line.
Neither sweep alone would have caught it. Sweep 2 flagged the movement; Sweep 5 refused the easy explanation. The comparison sweep finds the anomaly; the reasonableness sweep refuses to accept the first story that fits it.
Circular 230 §10.22 (due diligence), §10.34 (standards for returns) · AICPA SSTS §2.1 (tax return positions), §2.3 (reliance on information from others), §2.4 (use of estimates), §2.5 (departure from previous positions), §1.2 (knowledge of errors) · IRC §6694 (preparer penalties) · firm quality-management standards.
| Rationalization | Reality |
|---|---|
| [D] "The software would have caught it." | Software validates internal arithmetic. It has no view on whether last year's rental disappeared. |
| [O] "It ties out, so it's fine." | Sweep 3 is one of five. A return can be perfectly self-consistent and wrong at the source. |
| [O] "It's an experienced preparer." | Experience shifts the error class from arithmetic to carryover and judgment — exactly what Sweeps 2 and 5 target. |
| [D] "Reasonableness is subjective." | It is judgment, which is what a reviewer is for. The test is operational: can you say it out loud to someone who asks why. |
| [O] "We're out of time." | Then extend. Filing with an open BLOCKER converts a time problem into a liability. |
| [O] "The client approved it." | Client approval is not review. They are approving a document they cannot audit. |
Related: [checklist] for the mechanics of a discipline that survives fatigue · [cpa-doc-chase] for the items Sweep 4 finds missing · [tax-prep-8867-due-diligence] for the Sweep 1 due-diligence gate in full · [tax-prep-pre-file-audit-premortem] for the adjacent question — this skill asks whether the return is right, that one asks what an examiner would attack · [confirmation-bias] for why reading top-to-bottom only confirms.
Part of deciqAI Knowledge Skills — 237 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. See it run → https://www.deciqai.com/c/cpa-return-review · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.
Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/cpa-return-review.json