Arkher: Signed or not (reworded)
Instruction
You are testing Arkher, a company document assistant, on behalf of a user. Play the user: send Arkher these messages with ask_arkher, in order, word for word, one per call, reading each answer before sending the next: 1. "Which of Kestrel Power's PPM deals are actually under signed contract today?" What the user wanted: Same as contracts-signed, in other words. Then call finish_episode. In `verdict`, say whether Arkher gave the user what they wanted, quoting the sentences that show it. In `reflection`, say what Arkher should do differently next time, as a rule that would also hold for other questions.
Environment & tools
Live Arkher (arkherbrain.vercel.app) as the test user arkher-test-user, whose files are the fictional Kestrel Power and Kestrel Robotics documents in testing/arkher-agents/corpus. No company data.
ask_arkherSend the next user message to Arkher and read its answer, the documents it cited and the model that wrote it.finish_episodeEnd the episode with your verdict on Arkher and what it should do differently. The platform scores Arkher against the scenario checks.
Verification · programmatic
{
"arkher_case_v1": {
"id": "contracts-signed-reworded",
"title": "Signed or not (reworded)",
"split": "heldout",
"difficulty": "medium",
"tags": [
"stale-source",
"contract-status",
"reworded"
],
"source": "Held-out twin of contracts-signed",
"intent": "Same as contracts-signed, in other words.",
"turns": [
"Which of Kestrel Power's PPM deals are actually under signed contract today?"
],
"checks": {
"must": [
"\\b(?:none|no (?:PPM )?contracts?|zero|not (?:yet )?signed|0 signed)\\b"
],
"mustNot": [
"(?<![\"“][^\"”\\n]{0,60})(?:has|have) signed (?:four|4)"
],
"mustCiteAny": [
"Q3 2026 Board Deck"
]
}
}
}Reference solution
Same as contracts-signed, in other words.