01Safety
0.0 / 30
What changed in the harness
Selection accuracy 84→89, token cost down 7%, 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
26.9 / 30
03Economics
20.0 / 20
04Discoverability
12.4 / 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 |
|---|---|---|
need_human |
no_return_description |
Call when you cannot proceed because the next step requires a real human (Terms of Service or legal agreements, account creation needing identity verification/CAPTCHA, human-authenticated logins such as BankID/SMS 2FA/OAuth, or any step requiring legal personhood or physical presence). Returns the task_id of the dispatched task, which you then poll with check_task_status (typical completion 2-30 minutes; poll no more than once every 30 seconds). DO NOT use for sending messages, making HTTP requests, writing code, or any task you can complete with your existing tools. Format description as numbered steps, one instruction per line; put each URL on its own line; end with "REPLY WITH:" listing expected deliverables. Never include passwords, card numbers, CVCs, bank account numbers, SSNs, auth tokens, or API keys; for paid steps, tell the human what to pay for and where (they use their own payment method). Set demo:true for an instant synthetic response to verify integration without consuming credits or involving a real human. |
Selection evidence
Confusable tool pairs.
0 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.
| Tool A | Tool B | Confidence | Why they collide |
|---|---|---|---|
| No confusable tool pairs were flagged in this assessment. | |||
Compare the field