01Safety
30.0 / 30
What changed in the harness
Selection accuracy 93→96, token cost down 1%, 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
28.2 / 30
03Economics
16.8 / 20
04Discoverability
18.7 / 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.
1 defect 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 |
|---|---|---|
list_agents |
no_return_description |
Browse agents from the A2ASearch directory by category, optionally filtered by type ('MCP Server', 'CLI Tool', 'AI Coding Agent', 'Agent Skill', 'A2A Agent') and sorted by stars (most popular) or new (recently added). Returns a list of matching agents with name, description, type, stars and GitHub URL. |
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_agents |
list_agents |
medium | A task like 'show me top MCP servers' could be done via search_agents with a type filter or list_agents with type+sort; both accept type filters and overlap when the user wants a category browse rather than a keyword search. |
get_agent |
list_agents |
low | get_agent requires a specific slug while list_agents browses by category; confusion is unlikely unless the user names an agent ambiguously, but the tools serve clearly distinct purposes (single lookup vs browsing). |
Compare the field