Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Review and organize Todoist tasks in the Evolution project. Finds uncategorized, untranslated, or messy tasks and organizes them with proper categories, PT-BR translation, and actionable titles. Use when user says 'review todoist', 'organize tasks', 'todoist triage', 'clean up todoist', or similar.
.claude/skills/evolution-foundation-prod-review-todoist/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 1% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -3% | 0% |
Skill to review and organize tasks in the Evolution Todoist project. Identifies uncategorized, English-only, generic, or disorganized tasks and fixes them directly.
todoist installed and authenticatedEvolution project existing in Todoistbashtodoist tasks -p "Evolution"
A task needs triage if it meets any of these criteria:
[Categoria] prefix in the titleFor each task that needs triage, apply:
Add [Categoria] prefix to the title. Available categories:
| Category | When to use | |---|---| | [Produto & Tech] | Development, bugs, features, infra, deploy, code | | [Marketing] | Content, campaigns, videos, social media, launches | | [Sales] | Pipeline, proposals, partnerships, leads, pricing | | [Financial] | Accounts, invoices, payments, financial metrics | | [Operations] | Internal processes, groups, access, team communication | | [Strategy] | OKRs, roadmap, analyses, strategic decisions | | [Community] | Discord, support, user feedback, beta testers | | [Roadmap] | Future roadmap items, feature evaluation |
If the title is in English, translate to Brazilian Portuguese keeping clarity and objectivity.
Before: Send event registration link to team member After: [Operação] Enviar link de inscrição do evento para membro do time
The title should make clear:
Before: Upload web panel; grant team member access After: [Operação] Publicar painel web do evento e liberar acesso para membro do time
bashtodoist update <task-id> --content "[Categoria] Título traduzido e acionável"
Fundamental rule: execute the organization directly, without an intermediate report.
Do not list tasks before organizing. Do not ask for confirmation on each one. Organize all at once and confirm at the end.
Save a short report to workspace/daily-logs/[C] YYYY-MM-DD-todoist-review.md com:
markdown# Todoist Triage — YYYY-MM-DD **Project:** Evolution **Tasks reviewed:** {N} **Organized:** {M} (categorized, translated, or rewritten) **Already OK:** {K} (no changes needed) ## Organized Tasks | Task | Before | After | |--------|-------|--------| | ... | ... | ... |
Create the directory workspace/daily-logs/ if it does not exist.
When finished, present only:
## Todoist Triage — Completed
**Project:** Evolution
**Tasks reviewed:** {N}
**Organized:** {M} (categorized, translated, or rewritten)
**Already OK:** {K} (no changes needed)If the user wants to see details of what changed, they ask.
Evolution — do not move tasks to other projects[Category] and are in PT-BR)[Operations] as fallback| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,822 | 8,020 | +38% | 1 | 1 | 0% | 307 | 1,544 | +403% | 0 | 0 | — |
case-02 | fail→fail | 3,440 | 5,177 | +50% | 1 | 1 | 0% | 241 | 1,141 | +373% | 0 | 0 | — |
case-03 | fail→fail | 3,984 | 5,997 | +51% | 1 | 1 | 0% | 302 | 1,541 | +410% | 0 | 0 | — |
case-04 | fail→fail | 7,483 | 8,128 | +9% | 1 | 1 | 0% | 343 | 2,254 | +557% | 0 | 0 | — |
case-05 | fail→fail | 6,114 | 6,411 | +5% | 1 | 1 | 0% | 406 | 1,290 | +218% | 0 | 0 | — |
case-06 | fail→fail | 5,860 | 4,792 | -18% | 1 | 1 | 0% | 916 | 1,311 | +43% | 0 | 0 | — |
case-07 | fail→fail | 6,966 | 7,605 | +9% | 1 | 1 | 0% | 1,124 | 1,530 | +36% | 0 | 0 | — |
case-08 | fail→pass | 7,470 | 2,600 | -65% | 1 | 1 | 0% | 1,142 | 1,356 | +19% | 0 | 0 | — |
case-09 | fail→fail | 8,759 | 7,908 | -10% | 1 | 1 | 0% | 1,295 | 1,542 | +19% | 0 | 0 | — |
case-10 | fail→pass | 6,639 | 2,944 | -56% | 1 | 1 | 0% | 1,076 | 1,448 | +35% | 0 | 0 | — |
case-11 | fail→fail | 5,856 | 4,557 | -22% | 1 | 1 | 0% | 808 | 1,776 | +120% | 0 | 0 | — |
case-12 | fail→pass | 7,285 | 2,281 | -69% | 1 | 1 | 0% | 1,208 | 1,301 | +8% | 0 | 0 | — |
case-13 | fail→pass | 9,557 | 3,116 | -67% | 1 | 1 | 0% | 1,487 | 1,496 | +1% | 0 | 0 | — |
case-14 | fail→pass | 10,183 | 3,784 | -63% | 1 | 1 | 0% | 1,570 | 1,523 | -3% | 0 | 0 | — |
case-15 | pass→fail | 8,459 | 3,018 | -64% | 1 | 1 | 0% | 1,519 | 1,487 | -2% | 0 | 0 | — |
case-16 | fail→fail | 8,344 | 2,762 | -67% | 1 | 1 | 0% | 1,282 | 1,371 | +7% | 0 | 0 | — |
case-17 | pass→pass | 12,286 | 2,541 | -79% | 1 | 1 | 0% | 1,783 | 1,324 | -26% | 0 | 0 | — |
case-18 | fail→fail | 10,434 | 2,396 | -77% | 1 | 1 | 0% | 1,613 | 1,441 | -11% | 0 | 0 | — |
case-19 | fail→fail | 8,356 | 2,952 | -65% | 1 | 1 | 0% | 1,200 | 1,529 | +27% | 0 | 0 | — |
case-20 | fail→pass | 10,973 | 1,940 | -82% | 1 | 1 | 0% | 1,773 | 1,256 | -29% | 0 | 0 | — |
case-21 | pass→pass | 7,441 | 2,135 | -71% | 1 | 1 | 0% | 1,201 | 1,310 | +9% | 0 | 0 | — |
case-22 | fail→pass | 7,362 | 3,128 | -58% | 1 | 1 | 0% | 1,083 | 1,541 | +42% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 14 counted toward the lift figure. The other 8 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +27 percentage points is the difference between those two pass rates over the 14 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.