Trello

Audits for the Trello boards execution teams actually use under the official plan — the Card Aging Power-Up that resets on every modify, the Butler rules that fail silently, the archive that hides abandoned cards, the checklist items that drift unseen. Each entry ships with the tool, a user guide, and the prompts to build a similar one yourself. When the PMO asks "where's the work?", the board has to answer cleanly.

→ Entry 01 PowerShell · Local · RED/AMBER/GREEN verdict →

Card Aging Rebuilt

A local tool that replaces Trello's broken Card Aging Power-Up. Separates creation-age, in-list-age, and last-substantive-activity-age (the Power-Up conflates all three and resets on any modify, including a list-move). Tiers staleness by list semantics so a Blocked card stuck for 9 working days surfaces as RED even when nobody noticed. Bundled 42-card Apex Defense fixture across 8 lists runs end-to-end with no creds and produces a RED verdict via the "Firmware OTA rollout blocked" card.

→ Entry 02 PowerShell · Local · Five-verdict severity scale →

Archive Graveyard Audit

A local tool that walks every archived card on a Trello board and classifies it — was this archive a legitimate completion (Done list, checklist 100%, no open mentions) or a graveyard abandonment (Doing/Blocked archive with unfinished checklist, unresolved @mention, or due date still in the future)? Five-verdict severity scale (GRAVEYARD-CRITICAL through CLEAN-ARCHIVE) so the worst cases are unmissable. Bundled 38-card Apex Defense fixture engineers 5 GRAVEYARD-CRITICAL cards (3 abandoned from Doing, 2 from Blocked, 2 archived before their due date) to trigger the RED verdict.

→ Entry 03 PowerShell · Local · Four ghost verdicts · Card-health rollup →

Checklist Subtask Audit

A local tool that walks every checklist on every non-archived Trello card and surfaces the ghost items native views can't show: items assigned to deactivated members, items past their own due date, items on cards whose parent due date already passed, and checklists bloated past the 15-item anti-pattern threshold. Trello's board view exposes only the 3/7 aggregate badge — no filtering or reporting on individual checkItem state across cards. Bundled 44-checklist / 233-checkItem Apex Defense fixture engineers 4 GHOST-ORPHANED items and 1 severely-bloated 28-item checklist to trigger the RED verdict.

→ Entry 04 PowerShell · Local · Five verdicts · CRITICAL tiering →

Butler Health Report

A local tool that parses Trello's Butler automation-action stream, computes week-over-week fire cadence per rule, and surfaces the four silent-failure modes Butler's per-rule chronological log doesn't aggregate: DEAD (was firing, now zero), CHOKING (commandLimit ceiling hit repeatedly), DEGRADED (success ratio below 80%), STALE (no fires in 30+ days). CRITICAL tiering catches the intake/SLA/incident/escalation rules where silent failure is operationally most damaging. Bundled 24-rule Apex Defense fixture engineers 3 DEAD rules (2 CRITICAL) and 9 commandLimitExceeded hits to trigger the RED verdict three independent ways.

→ Cross-cluster A tool for reconciling Trello with an MS Project plan lives in the Microsoft Project cluster. If you're the execution team running this board under an official MSP plan, that's the audit you want. And the MSP ↔ Trello Sync — also in the Microsoft Project cluster — wraps the diff into operations: push the MSP plan to Trello before the status meeting, pull the changeset back after. Same matcher, same flag-both conflict policy, same don't-write-to-MSP discipline.

Claude could write this. Two reasons to prefer the script.

Some workplaces — defense, medical devices, finance, anywhere with a strict IT policy — won't let Claude or any AI tool touch production data. These tools run on your machine, with no Claude or AI in the data path — your Trello board data flows direct from your machine to api.trello.com/1 and back, no third-party SaaS in between, no telemetry, no tokens leaving your environment. And even where AI is allowed, repeat workflows shouldn't cost tokens — a deterministic script runs the same way every time, for free, forever. The prompt guide on each entry shows how the tool was built with Claude; the downloads are what you run after.