Menú
Human-readable status report for all tax forms. Shows benchmark accuracy, active root causes, and build/fix state for each form:year.
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Basado en la clasificación ocupacional SOC
Open-source tax engine. Onboards users, installs the CLI, and routes to the right skill -- preparing a return or reviewing one.
Tax preparer agent that uses the opentax CLI to prepare, validate, and export federal tax returns. Handles W-2s, 1099s, all major schedules, credits, and more.
Tax return reviewer that audits a completed return against source documents using the opentax CLI. Finds discrepancies, missing income, incorrect deductions, and missed credits.
Downloads and installs the opentax CLI binary. Detects OS and architecture automatically.
Helps users permanently set up the OpenTax skill on their AI platform so they don't have to load it every time.
Generic audit/fix loop. Pass a goal string. Planner subdivides work, auditor/fixer teams run in parallel, loop until 2 consecutive clean cycles.
| name | tax-status |
| description | Human-readable status report for all tax forms. Shows benchmark accuracy, active root causes, and build/fix state for each form:year. |
Glob forms/f*/FORM.md to enumerate all known forms.
Read .state/bench/state.json. If it does not exist → print "Harness not initialized." and stop.
Confirm version === 2. If version === 1 → print "State schema outdated. Run migration." and stop.
For each entry in state.forms, print one section:
━━━ Tax Harness Status ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
{form}:{year} Fix
Status: {fix.status}
Baseline: {fix.benchmark_baseline.pass} pass / {fix.benchmark_baseline.fail} fail
Progress: ████████░░ {pct}%
Root Causes:
✓ {name} (fixed)
✗ {name} (pending — cases {list})
{form}:{year} Build
Status: {build.status}
Phase: {build.phase}
Cases: {build.benchmark_cases_created} benchmark cases
Nodes: {build.nodes_built.length} nodes built
Last run: {last benchmark_runs entry}
━━━ Recent Activity ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
{last 5 entries from .state/bench/progress.md}
Use color: green for done/fixed, red for pending/failed, yellow for running.