Compressed caveman-style prose for AI coding agents — cuts ~65% output tokens while keeping full technical accuracy
96
100%
Does it follow best practices?
Impact
96%
1.00xAverage score across 38 eval scenarios
Passed
No known issues
{
"context": "Tests whether the response correctly maps application types to appropriate OAuth2 flows.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Server-side app: Authorization Code flow",
"description": "Recommends Authorization Code flow (with client secret) for the server-side web app since it can securely store the client secret",
"max_score": 12
},
{
"name": "SPA: Authorization Code with PKCE",
"description": "Recommends Authorization Code flow with PKCE for the SPA since it cannot store a client secret (NOT the deprecated Implicit flow)",
"max_score": 12
},
{
"name": "CLI: Device Authorization flow",
"description": "Recommends Device Authorization flow (or Authorization Code with localhost redirect) for the CLI tool",
"max_score": 12
},
{
"name": "Explains security rationale",
"description": "Explains why different flows exist: client secret confidentiality, redirect URI validation, token exposure in browser",
"max_score": 10
},
{
"name": "No incorrect information",
"description": "OAuth2 flow names, mechanics, and security properties are technically correct",
"max_score": 12
}
]
}evals
scenario-1
scenario-2
scenario-3
scenario-4
scenario-5
scenario-6
scenario-7
scenario-8
scenario-9
scenario-10
scenario-11
scenario-12
scenario-13
scenario-14
scenario-15
scenario-16
scenario-17
scenario-18
scenario-19
scenario-20
scenario-21
scenario-22
scenario-23
scenario-24
scenario-25
scenario-26
scenario-27
scenario-28
scenario-29
scenario-30
scenario-31
scenario-32
scenario-33
scenario-34
scenario-35
scenario-36
scenario-37