cpa-doc-chase
Skill13 starsChanged 4 months agoMITView on GitHub
Activate when: a tax or bookkeeping engagement is blocked on client paperwork; a client has gone quiet on a document request; drafting an initial PBC/document request; deciding whether to extend, deliver incomplete, or disengage; user says 'chase the client', 'missing documents', 'PBC list', 'client won't respond', 'what am I still waiting on?'. Do NOT activate when: the documents are in hand and the question is technical treatment; the request is to prepare or review the return itself. More: deciqai.com/s/cpa-doc-chase
In this repository252 other files it gives its agents.
- 10-10-10Skill
- aarrr-pirate-metricsSkill
- abductive-reasoningSkill
- activation-onboarding-playbookSkill
- aeo-answer-engine-optimizationSkill
- agent-delegability-assessmentSkill
- agent-ops-monitoringSkill
- agent-outreach-lead-nurturingSkill
- agent-ready-product-dataSkill
- ai-content-factory-brand-consistencySkill
- ai-disruption-risk-auditSkill
- ai-output-verification-trust-tiersSkill
- ai-pilot-to-scaleSkill
- ai-risk-compliance-baselineSkill
- ai-roi-business-caseSkill
- anchoringSkill
- ansoff-matrixSkill
- antifragileSkill
- ar-dso-disciplineSkill
- arrow-information-paradoxSkill
- authority-biasSkill
- availability-heuristicSkill
- batna-zopaSkill
- bayesian-reasoningSkill
- bcg-matrixSkill
- black-swanSkill
- blue-ocean-strategySkill
- bootstrapping-vs-raisingSkill
- build-buy-partner-aiSkill
- bullseye-traction-channelsSkill
- business-model-canvasSkill
- cash-conversion-cycleSkill
- change-and-constantsSkill
- checklistSkill
- chestertons-fenceSkill
- circle-of-competenceSkill
- cognitive-evolution-stagesSkill
- cognitive-load-theorySkill
- cognitive-science-landscapeSkill
- commitment-consistencySkill
- comparative-advantageSkill
- compound-interestSkill
- confirmation-biasSkill
- content-led-growthSkill
- continuous-discoverySkill
- contrarian-questionSkill
- cpa-form-finderSkill
- cpa-irs-representationSkill
- cpa-return-reviewSkill
- critical-thinkingSkill
- cross-disciplinary-creativitySkill
- curiosity-learning-curveSkill
- customer-relationship-ladderSkill
- customs-adcvd-tariff-exposureSkill
- customs-duty-optimization-opportunity-costSkill
- customs-hts-classification-meceSkill
- customs-origin-usmca-decisionSkill
- customs-reasonable-care-isf-checklistSkill
- cynefinSkill
- death-spiralSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 92
Every item is a BLOCKER. Nobody triaged.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
CPA — Client Document Chase
Not legal or tax advice. Verify required documentation against your engagement letter, firm standards, and current IRS/state guidance. No penalty rates, retention periods, extension effects or plan-funding deadlines appear here — all of them expire. Route every such question to current authority rather than supplying a remembered figure, including inside a message drafted for a client. The practitioner who signs owns the file.
Activate when: an engagement cannot progress without client materials; a client has gone silent; you are drafting a document request or a follow-up; a deadline is approaching and the chase has not converged; you are handing an engagement to a colleague and need the blocking state written down. Do NOT activate when: the materials are complete; the question is how to treat an item you already have.
Why this skill
Engagements rarely stall on technical difficulty. They stall on paperwork that never arrives, and then on follow-up that is polite, unstructured, and ignored.
The specific failure: a flat alphabetical list of twenty missing items produces either paralysis or delivery of the five easiest. Neither unblocks you. The fix is to sort by what an item blocks, not by what it is, and to escalate on a fixed clock rather than on how the preparer feels that week.
Process
1. Frame the engagement. Engagement type and year · entity types and jurisdictions · prior-year working papers available · hard deadline and whether extension exists · who actually holds the documents (often a bookkeeper or spouse, not the signer) · preferred channel and past response pattern · fee status. Mark anything unknown as UNKNOWN — an unknown here is itself a finding.
Gate: no prior-year file and no engagement letter → you are scoping, not chasing. Scope first.
2. Mine the prior year. Walk last year's return or working papers line by line: what document supported this, and do I have it yet? This finds what a generic checklist never will — the K-1 from a partnership the client forgot they hold, the state they moved out of mid-year.
3. Build the item list. For each: what it is in the client's language · one-sentence why · class · source institution · status · substitute.
| Class | Meaning |
| BLOCKER | Cannot start or cannot finish without it |
| COMPLETENESS | Can proceed; result is not signed off until it lands |
| NICE-TO-HAVE | Improves outcome, does not gate delivery |
The substitute column is the one most firms skip and the one that resolves the most stuck items. A client who cannot find a document can often produce a statement, portal export, or letter serving the same evidentiary purpose. Naming it up front converts "I can't find it" from a dead end into a second option.
The why column decides whether anything arrives. "Required for the return" produces nothing. "Without this we can't support the deduction if it's questioned, so we'd have to leave it off" produces documents.
Gate: more than ~10 BLOCKERs → the engagement is mis-scoped or the client is not ready. Escalate to a scoping conversation, not a longer list.
4. Write the request. Five properties, all required:
- Short — blockers only, under ten items, ideally under five
- Specific — "your [year] mortgage interest statement from [lender]", not "mortgage documents"
- Dated — a real date with working room before the true deadline, never "at your earliest convenience"
- Consequence stated once, neutrally — extension, delay, or item excluded. A fact, not a threat, and not repeated
- One action — one channel, one reply, one destination. Never offer three upload options
Then cut it by a third. Length reads as difficulty.
5. Run the ladder. Fixed intervals, rising specificity, changing channel. Never send the same message twice.
| Stage | Timing | Change from previous |
| 1 · Request | Day 0 | Blockers only, dated, one channel |
| 2 · Reminder | +5–7d | Same list, shorter; lead with what's outstanding |
| 3 · Narrow | +5d | Cut to the two or three that matter; offer substitutes |
| 4 · Channel change | +3d | Email → phone, or reach the actual document holder |
| 5 · Decision notice | Deadline − buffer | State the decision and the date you make it |
| 6 · Act | At the date | Extend, deliver documented-incomplete, or disengage |
Gate: never restart the ladder silently. Partial delivery then silence = you are at stage 3 for the remainder, not back at stage 1.
Gate: stage 6 must actually happen. A ladder with no terminal action teaches the client that dates are decorative, and next year is worse.
6. Decide and record. Proceed complete · proceed with substitutes (record what was accepted for what, and why it was adequate) · proceed incomplete (record item, impact, client instruction, acknowledgement) · extend or disengage (record reason and date).
The chase log is the evidence — what was asked, when, how often, what the client said. If a delivered engagement is ever questioned, that record is the difference between a defensible position and a recollection.
Worked example
Rental-property client, filing deadline six weeks out. Twenty-two open items on the preparer's list; three rounds of "just checking in" have produced nothing.
Reclassified: 3 BLOCKERs (closing statement on the property sold in March, the K-1 from the client's brother's LLC, the mortgage interest statement on the refinanced loan), 11 COMPLETENESS, 8 NICE-TO-HAVE.
Stage 1 asked for the three, with substitutes named for two. The closing statement arrived in a day. The K-1 had not been issued at all — a fact the twenty-two-item list had been hiding for a month, and one that changed the plan from "chase" to "extend". Stage 5 went out on the same day the discovery was made, not six weeks later.
The finding was not a missing document. It was a missing document that could not exist yet. Ranking by blocker surfaced it; the flat list buried it.
Compliance anchors
Circular 230 §10.22 (due diligence) · §10.34(d) (reliance on client information) · engagement letter scope and termination terms · firm documentation retention policy · AICPA SSTS §2.3 (reliance on information from others).
Packs
- Solo: one ladder per client in the calendar the day the request goes out. Nothing else. The calendar entry is the whole system.
- Firm: blocker counts by preparer on the weekly stand-up; stage-5 notices reviewed by a partner before sending; the disposition record required before an engagement can be marked delivered.
Red flags
- The list has grown between rounds instead of shrinking — the preparer is adding items rather than escalating.
- Every item is a BLOCKER. Nobody triaged; they relabeled.
- The client responds warmly and sends nothing, repeatedly. Warmth is not progress; count documents.
- The document holder has never been contacted directly.
- Stage 5 has been sent twice. It cannot be — the second one is stage 1 of a fiction.
- An unpaid balance is being managed by not mentioning it, while work continues.
Common Rationalizations
| Rationalization | Reality |
| [O] "I don't want to be pushy — they're a good client." | Good clients are the ones who can absorb a dated request. The chase you avoid becomes the extension they didn't want. |
| [D] "I'll send the whole list so they can do it all at once." | Twenty items reads as a project and gets scheduled for never. Three reads as a task. |
| [O] "They said they'd send it this week." | Promised is a status, not a receipt. The ladder does not pause for a promise. |
| [O] "It's faster to just call them again." | Repeating stage 2 in a new medium is still stage 2. Escalation means changing the ask, not the channel alone. |
| [D] "We'll extend anyway, so the date doesn't matter." | Then say so at stage 1 and set the real date. A soft date the client can sense is soft trains next year's behavior. |
| [O] "If I disengage I lose the fee." | An engagement that cannot be completed and cannot be documented is a liability carrying a receivable. |
| [O] "While I'm explaining the extension I'll just mention what the late-filing penalty runs and which plan deadlines it moves." | Observed: a remembered penalty ratio and a remembered plan-funding deadline, neither of them checked, both landing in a message addressed to the client. The chase plan is not the place to answer a rate or a date — name the question and route it. |
Verification
- Every item classed BLOCKER / COMPLETENESS / NICE-TO-HAVE
- Prior-year file walked line by line, not just a template list used
- Substitutes named for every blocker
- Initial request is blockers only, dated, single-channel
- Follow-up dates in the calendar the same day the request went out
- Actual document holder identified, not assumed to be the signer
- Terminal action taken at the stated date, not deferred
- Disposition and rationale in the engagement file in writing
- Every penalty rate, deadline and retention period routed to current authority, not answered from memory — checked against the client-facing messages too, not only the plan
Related: [checklist] for the standing item list · [reciprocity] and [door-in-the-face] for why the narrowed stage-3 ask outperforms the full list · [switching-costs] for why chronic non-responders are expensive to keep.
Part of deciqAI Knowledge Skills — 253 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/s/cpa-doc-chase · Built by deciqAI · github.com/deciqAI · Contributions welcome.
Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/cpa-doc-chase.json
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

