Skip to content

Loop XXI · Tools for agents

Useful work.
Ready for your agent.

Start with a local tool that turns an inquiry CSV into a clear follow-up queue.

Follow-up Worklist · Free download

Find the gaps
before work gets lost.

Flag overdue items, missing owners, unclear next actions and duplicate references. Keep every decision with the person responsible for the work.

Use it directly or give the included instructions to an agent that supports local skills. Your records stay in the local tool; the calling agent can see its output.

Download the agent kit ↗

Version 1.0.0 · Node.js 22+ · No account or credits

A synthetic example

Six inquiries.
A visible next step.

Total records
6
Need review
4
Closed, excluded
1

The example preserves both duplicate records, flags an overdue item and identifies an inquiry missing an owner, next action and due date. One scheduled record stays in the queue.

These are synthetic records, not customer results or measured time savings.

Try the browser version ↗

For agents and implementers

Inspect. Run. Review.

  1. Download the kit, compare its checksum and inspect the included files.
  2. Unzip into an empty folder. Run the synthetic example from the loop-worklist directory.
  3. Supply an authorized CSV and a review date. Review the returned flags before deciding what to do.
node scripts/worklist.mjs --date 2026-09-09 < assets/example.csv

Requires a local command runner. No dependency installation is needed. The command works directly; skill installation follows your agent host’s instructions.

A clear interface.

Six input columns. JSON results, an optional spreadsheet-safe CSV export, explicit errors and a hash linking the result to the supplied input.

Verify the download

Compare the archive’s SHA-256 with this value and the manifest. File checksums are also included inside the kit.

7f2568263ff43cd624c4edfad2d40ed897ec3adc349d10cf1b27ada5784831aa

Checksums identify bytes. They are not independent security certification or proof that customer work was completed.

Built around review

Your workflow.
Your decisions.

The tool runs fixed review rules. It makes no AI calls, sends no messages, changes no business systems and takes no payments. Every output remains a review queue.

Local use has no telemetry. Download counts cannot tell us whether your task succeeded. Use only data your agent is authorized to handle.

Connect the next step.

The existing Inquiry Workflow Pilot adapts one intake-to-review workflow to customer-owned systems, with tested exceptions, acceptance evidence and a portable handoff.

Read the delivery approach ↗

The $750 pilot is inquiry-only, subject to written scope and availability. This download creates no order. Hosted AI credit products remain subject to their published availability.

AI infrastructure status ↗