01Safety
30.0 / 30
What changed in the harness
Selection accuracy 93%, destructive-action safety rate 100% (baseline only -- no rewrite pass applied).
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
27.2 / 30
03Economics
17.0 / 20
04Discoverability
17.9 / 20
Highest-impact fix
Estimated gain +3 pointsReduce the tool-definition and task token payload
Consolidate overlapping tools and trim repetitive schema text while preserving behavior, parameter, and return semantics.
Description evidence
Defects and rewrites.
0 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 |
|---|---|---|
| No description defects were flagged in this assessment. | ||
Selection evidence
Confusable tool pairs.
10 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.
| Tool A | Tool B | Confidence | Why they collide |
|---|---|---|---|
sakura_now |
sakura_forecast |
medium | Both tools speak to 'the sakura forecast'; sakura_now lists 'How is the sakura forecast?' as an example while sakura_forecast handles timing/peak dates. A broad forecast query without a status-vs-timing cue can be routed to either. |
koyo_now |
koyo_forecast |
medium | 'Kyoto autumn leaves forecast' is an explicit example in koyo_now while koyo_forecast is the dedicated timing/peak tool; a foliage forecast request that doesn't clarify 'current status' vs 'when it peaks' is ambiguous. |
fruit_seasons |
fruit_farms |
medium | Both descriptions mention 'fruit picking'; fruit_seasons covers best picking months/trips while fruit_farms covers actual farm listings and bookings. A picking request without explicit 'farm/booking/coordinates' wording is ambiguous. |
sakura_now |
sakura_best_dates |
low | Both accept travel dates and answer 'where to see sakura'; sakura_now returns current status/spots for the trip while sakura_best_dates matches windows, so a dated 'where should I view sakura' request could go either way. |
koyo_now |
koyo_best_dates |
low | Both take trip dates and answer 'where for fall foliage'; koyo_now gives current status and spot suggestions while koyo_best_dates matches cities to the date window, leaving a dated foliage inquiry ambiguous. |
sakura_best_dates |
koyo_best_dates |
low | Structurally identical dateline-matching tools, and the season discriminator is the user's word choice; a travel-dates question that omits the season term could be routed to the wrong one. |
sakura_now |
sakura_spots |
low | 'Where should I view sakura' returns spot suggestions from sakura_now, while sakura_spots is the named per-prefecture spot tool; location-focused prompts without an explicit prefecture can select either. |
koyo_now |
koyo_spots |
low | Both answer 'where to see fall foliage'; koyo_now returns current good spots and koyo_spots the exact prefecture list, so a location-focused query is plausible in either tool. |
sakura_forecast |
sakura_spots |
low | A 'best sakura in Tokyo' request without stating whether peak dates or exact exact spots are needed could land on sakura_forecast (city filter) or sakura_spots (prefecture spots), despite the explicit cross-references. |
koyo_forecast |
koyo_spots |
low | A 'best koyo in Nikko' request with only a region given could be routed to koyo_forecast (region filter) instead of koyo_spots (exact spots), though the descriptions do cross-reference each other. |
Compare the field