01Safety
0.0 / 30
What changed in the harness
Selection accuracy 98→94, token cost up 1%, unconfirmed writes 0%→0%.
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
29.1 / 30
03Economics
20.0 / 20
04Discoverability
12.6 / 20
Highest-impact fix
Estimated gain +30 pointsAdd explicit identity and permission preflight tools
Expose machine-readable principal/tenant confirmation and a non-mutating permission check so agents can verify both before destructive actions.
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_checks |
no_return_description |
Return the full catalog of checks the scanner runs as a list of check objects. Each returned check includes its id, name, category (Cloudflare isitagentready-style), a description of what the check verifies, and a link to the relevant specification. |
Selection evidence
Confusable tool pairs.
1 pair where similar names or overlapping descriptions may send an agent toward the wrong tool.
| Tool A | Tool B | Confidence | Why they collide |
|---|---|---|---|
get_fix_plan |
get_aso_framework |
medium | Both descriptions revolve around ASO scoring of signals: get_fix_plan returns 'the ASO Score points each fix is worth' and artifacts per signal, while get_aso_framework returns 'six signal pillars with point weights' and the scoring rubric. A natural-language ask like 'what are the ASO point values / how is my ASO score weighted / how many points is a signal worth' maps plausibly to either tool, since the generic scoring-rubric content of get_aso_framework and the per-fix point values of get_fix_plan overlap and neither clearly gates conceptual scoring questions. |
Compare the field