Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Naming conventions for artifact kind and path, ensuring data can be aggregated and analyzed
.claude/skills/chekusu-artifact-naming/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 82% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 64% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -23% | 0% |
Every call to wanman artifact put must include the --path parameter. Artifacts without a path cannot be retrieved or aggregated.
bash# Full command format wanman artifact put --kind <kind> --path "<domain>/<category>/<item>" --source "<source>" --confidence <0-1> '<json>'
After writing an MD/HTML or other file, use the --file parameter to store the file contents in the artifact. This ensures file contents are version-controlled in the database.
bash# After writing a file, include --file in artifact put wanman artifact put --kind brand_asset --path "brand/identity/handbook" \ --source "marketing" --confidence 0.9 \ --file /workspace/agents/output/marketing/brand-design.md \ '{"name":"Brand Handbook","type":"handbook"}' # Structured data (no file) does not need --file wanman artifact put --kind budget_item --path "costs/opex/rent" \ --source "estimate" --confidence 0.4 \ '{"item":"rent","amount":350000,"currency":"JPY"}'
Rule: If your task output includes MD or HTML files, you must include --file pointing to that file when calling artifact put.
Only use the following standard kind values — do not invent your own:
| Domain | Standard Kind | Description | |--------|--------------|-------------| | Market research | competitor | Competitor information | | Market research | market_data | Market data (size, trends) | | Market research | customer_profile | Customer personas | | Market research | location_data | Location-related data | | Finance | budget_item | Individual cost item (use this uniformly, not monthly_cost/initial_investment) | | Finance | revenue_forecast | Revenue forecast | | Finance | financial_summary | Financial summary (break-even, payback period, etc.) | | Branding | brand_asset | Brand asset (name, story, visuals, tagline — all under brand_asset) | | Content | content_plan | Content planning | | Technical | tech_spec | Technical specification |
Format: {domain}/{category}/{item}
costs/capex/renovation ← One-time renovation cost
costs/capex/equipment ← Equipment procurement
costs/opex/rent ← Monthly rent
costs/opex/labor ← Monthly labor
costs/opex/materials ← Raw material costs
market/competitors/cafe-a ← Competitor A
market/competitors/cafe-b ← Competitor B
market/demographics/target ← Target customer segment
market/location/rent-level ← Rent level
brand/naming/candidate-1 ← Brand name candidate 1
brand/naming/candidate-2 ← Brand name candidate 2
brand/visual/colors ← Color system
brand/visual/typography ← Typography direction
brand/story/core ← Brand story
brand/slogan/main ← Brand tagline
content/social/instagram ← Instagram plan
content/social/twitter ← Twitter plan
revenue/forecast/monthly ← Monthly revenue forecast
revenue/forecast/breakeven ← Break-even analysis
finance/summary/cashflow ← Cash flow summary
tech/website/spec ← Website technical specificationbash# Wrong: kind is too granular, no path wanman artifact put --kind initial_investment --source "estimate" --confidence 0.4 '{"item":"equipment"}' wanman artifact put --kind monthly_cost --source "estimate" --confidence 0.4 '{"item":"rent"}' wanman artifact put --kind brand_name_candidate --source "original" --confidence 0.8 '{"name":"ITTEKI"}' # Correct: standard kind + path for differentiation wanman artifact put --kind budget_item --path "costs/capex/equipment" --source "estimate" --confidence 0.4 '{"item":"equipment","amount":500000}' wanman artifact put --kind budget_item --path "costs/opex/rent" --source "estimate" --confidence 0.4 '{"item":"rent","amount":350000}' wanman artifact put --kind brand_asset --path "brand/naming/candidate-1" --source "original" --confidence 0.8 '{"name":"ITTEKI","meaning":"one drop"}'
Other measured skills in the registry, with their headline benchmark lift.