01Safety
30.0 / 30
What changed in the harness
Selection accuracy 100%, destructive-action safety rate 100% (baseline only -- no rewrite pass applied).
Category breakdown
Earned points across the four signals Gradable measures. Safety and Legibility are scored out of 30; Economics and Discoverability are scored out of 20.
30.0 / 30
29.2 / 30
20.0 / 20
20.0 / 20
Highest-impact fix
Estimated gain +1 pointGive confusable tools explicit decision boundaries and examples of when to choose one instead of the other.
Description evidence
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
3 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.
| Tool A | Tool B | Confidence | Why they collide |
|---|---|---|---|
get_item_details |
get_item_price_history |
high | Both descriptions are centered on item price and price trends; examples overlap (details lists percent price change over 30/90/180 days, history lists daily prices for 180 days for trends). A task such as what is the price trend of a dragon scimitar or simply the price of an item without the word history could route to either tool. |
get_player_count |
get_rsuser_total |
medium | Both take no input and report a count about RuneScape: players currently online vs total accounts ever created. A task like how many RuneScape players are there or tell me the number of accounts is ambiguous between the two measures. |
get_player_hiscore |
get_top_rankings |
low | Both return rank/level/score data for players in a skill or activity and share the concept of rankings. A task like what is the top ranking in agility or compare player rankings in an activity could plausibly map to either, although the required playerName in hiscore usually disambiguates a named-player query. |
Compare the field