Use when connecting a workflow to Discord using the API
85
90%
Does it follow best practices?
Impact
69%
1.01xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
{
"context": "Tests whether the agent produces a properly structured Discord Post Draft in the correct format and writes community-channel-appropriate copy: calm, clear, friendly, non-defensive, with short paragraphs, light bullets, and minimal jargon.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Has ## Draft section",
"description": "Output file contains a '## Draft' heading",
"max_score": 8
},
{
"name": "Has ## Notes section",
"description": "Output file contains a '## Notes' heading with at least one item",
"max_score": 8
},
{
"name": "Short paragraphs",
"description": "No prose paragraph in the Draft section exceeds 4 lines of text",
"max_score": 10
},
{
"name": "Uses bullets",
"description": "The Draft section uses bullet points (- or •) for listing release changes rather than a dense prose block",
"max_score": 10
},
{
"name": "Minimal jargon",
"description": "The Draft does NOT use raw technical terms without explanation — e.g., 'inotify', 'FSEvents', 'lazy-loading plugin registry', or 'race condition' appear only if explained in plain language",
"max_score": 12
},
{
"name": "Non-defensive tone",
"description": "The Draft does NOT contain defensive or apologetic framing around the sync bug fix (e.g., does not say 'we messed up' or over-explain the corruption issue in alarming terms)",
"max_score": 12
},
{
"name": "Calm and friendly tone",
"description": "The Draft reads as calm and friendly — no all-caps urgency, no alarming language, no corporate defensiveness",
"max_score": 10
},
{
"name": "Known issue disclosed",
"description": "The Draft or Notes mentions the known issue with --watch mode and symlinked directories",
"max_score": 10
},
{
"name": "Breaking change flagged",
"description": "The removal of --legacy-auth is mentioned in the Draft or Notes as a breaking change or migration note",
"max_score": 10
},
{
"name": "No invented changes",
"description": "The Draft does not describe features, fixes, or improvements that are not present in the input notes",
"max_score": 10
}
]
}