01Safety
30.0 / 30
What changed in the harness
Selection accuracy 98→98, token cost up 4%, unconfirmed writes 100%→100%.
Category breakdown
Where the score comes from.
Earned points across the four signals Gradable measures. Safety and Legibility are scored out of 30; Economics and Discoverability are scored out of 20.
02Legibility
27.9 / 30
03Economics
20.0 / 20
04Discoverability
18.4 / 20
Highest-impact fix
Estimated gain +1 pointRewrite 5 underspecified tool descriptions
State each tool's behavior, explain every parameter, and declare return semantics, prioritizing tools exercised by failed selection tasks.
Description evidence
Defects and rewrites.
5 defects found across the exposed tool descriptions. Suggested rewrites make purpose, inputs, boundaries, and returns easier for an agent to understand.
| Tool | Defect types | Suggested rewrite |
|---|---|---|
browse_sets |
no_return_description |
Browse One Piece TCG sets. If set_name is omitted, returns a list of all sets with card counts; if set_name is provided, returns the cards in that set (up to limit) with their details. |
card_versions |
no_return_description |
Look up every printing/version of a One Piece TCG card, including alternate art variants, by card name or card number. Returns a list of matching printings with their set, number, and attributes. |
browse_leaders |
no_return_description |
List all One Piece TCG Leader cards. Optionally filter by color to find leaders for a specific deck color. Returns matching leader cards with their details. |
analyze_cost_curve |
no_return_description |
Analyze the cost curve of a One Piece TCG deck list. Returns cost distribution, color breakdown, and card type spread for the deck provided in deck_list. |
find_counters |
no_return_description |
Find One Piece TCG cards with counter values, useful for building defensive deck strategies. Optionally filter by color and maximum cost. Returns matching cards with their counter values and details. |
Selection evidence
Confusable tool pairs.
2 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.
| Tool A | Tool B | Confidence | Why they collide |
|---|---|---|---|
search_cards |
card_versions |
medium | Both tools are card-name lookups; a task like 'find all versions/printings of Monkey.D.Luffy' could be routed to search_cards (generic name search returning card details) instead of card_versions, which is the dedicated printings/alternate-art tool, so an agent may pick the wrong member. |
browse_sets |
browse_leaders |
low | A task like 'browse the leaders in set OP-04' is weakly ambiguous: browse_leaders matches 'leaders' but has no set filter, while browse_sets can show cards within a set (including leaders), so an agent could waver between the two, though the nouns 'sets' vs 'leaders' usually disambiguate. |
Compare the field