device_serial string | note string | note2 string | items list | note3 string |
|---|---|---|---|---|
RS7XKZDI8HTOJNYL | CONTACT POLICY (2026-08-01, user-mandated): any task that touches contacts/messages/calls/groups may ONLY use these 4 REAL contacts — Yuvraj Singh Jio (+919354672378), Yuvraj Airtel (+919354672378, = the device's own number), Maa (+918130285662), Dad (+919560156082). All earlier fabricated contacts (DailyBench Test, Ro... | NOTES APP MIGRATION (2026-08-02, user chose 'Mixed: Obsidian + keep Notes'): note-CREATING tasks (hard 1/4/7/8/10 + Calculator medium 'Sales Tax vs Flat Fee') now target OBSIDIAN (md.obsidian) — notes are plain .md files in the vault at /sdcard/Obsidian/Papers vault oneplus / (trailing space in dir name), so seeding/ve... | [
{
"placeholder": "[contact]",
"created": false,
"what": "RESOLVED to the REAL safe contact 'Yuvraj Airtel' (+919354672378 = the device's own number) per the 2026-08-01 user policy. The earlier fabricated 'DailyBench Test' contact was DELETED.",
"created_at_utc": null,
"identifiers": null,
"d... | PUBLIC SAMPLE SEEDING (2026-08-18): to make the rebuilt public.md (a true 530 sample) fully solvable, 4 entities were seeded on-device: Obsidian 'Shared Bill' note, a 120MB seed file, a 'Daily Reflection' Notes-app note, and a 'Q3 Review' Slides presentation. All recorded above + in the per-task manifests (assets/seeds... |
DrainBench-530 — Android agent benchmark (real phone, real LLM)
DrainBench runs Android agent tasks against a real phone (via ADB/mobilerun) and a real LLM, and grades the agent on reaching a verifiable device end-state. This repo ships the 530-task corpus plus everything needed to reproduce runs.
The 61-task public 3-day sample (and its personal/device-specific vars + seeds) is kept in a separate private repo; this public repo carries only the 530 corpus + its run-time variables + fabrication disclosure.
Files
| File | Content |
|---|---|
data/DailyBench_530_v1.json / .jsonl |
The 530-task corpus (Easy 1pt / Medium 3pt / Hard 5pt). |
data/tasks_530.md |
Human-readable source (canonical prompt text). |
data/tasks.md |
Wider task list (superset). |
data/multiturn_kb_530.json |
Knowledge-base profiles for ASK USER - MULTI tasks. |
data/ask_user_facts.json |
Facts for ASK USER SINGLE tasks. |
data/hallucination_controls.json |
Hallucination-control tasks. |
vars/ |
530-corpus run-time variable values (resolve every [placeholder]). |
config/user_config.example |
Documented default persona/device config. |
fabrication/ |
Full disclosure of every fabricated test entity. |
Task model
- DETERMINISTIC — everything needed is seeded on-device; end state is ADB-verified.
- ASK USER SINGLE — one fact deliberately omitted; agent must ask.
- ASK USER - MULTI — multi-turn dialogue against the KB profile.
Each record carries task_id, exact prompt_text/prompt_template,
placeholder slots, difficulty, and grading type. Prompts use [placeholder]
slots resolved through the run-time variables in vars/.
Fabricated test data — disclosure
To make deterministic tasks solvable, a controlled fictional persona and
fabricated data are seeded on the device. Everything is disclosed honestly in
fabrication/ (machine-readable JSON + full human-readable doc). No real phone
numbers, personal emails, or real identities appear; all personas are fictional.
- Downloads last month
- 21