Bread Agent Harness

Header baseline human_chrome_macos has no captured-header session in the last 30 days — the Header anomalies column is suppressed (shown as n/a, not 0). Run a human_chrome_macos reference session to restore the comparison.

Tags tracked

Cohort ∪ R2.

R2 sessions (7d)

Distinct sessionIds, all tracked tags.

Verified runs

results.jsonl rows with verified=true.

Flagged runs

Verified but operator notes non-empty.

Drift tags

In R2, not in any cohort JSONL.

Overview

One row per tag. Click a tag to inspect its identity, init signatures, and headers. R2 lookback: last 7 days (tag detail uses the wider window).

Tag	Dispatch	Header anomalies	Sessions in R2	Batches	Flagged	Verified	Init Rate	Mean Dur (s)	Mean Events	First Seen	Last Seen
Scripted
`scripted_playwright_vanilla_headless` compare ↔	Scripted	n/a	0	0	0	7	88%	0.0	0.0	—	—
`scripted_puppeteer_stealth_headless` compare ↔	Scripted	n/a	0	0	0	10	100%	0.0	0.0	—	—
`scripted_puppeteer_vanilla_headless` compare ↔	Scripted	n/a	0	0	0	21	81%	0.0	0.0	—	—
Stagehand auto
`agent_stagehand_claude_sonnet_4_6` compare ↔	Stagehand auto	n/a	0	0	0	3	100%	0.0	0.0	—	—
`agent_stagehand_gpt_5` compare ↔	Stagehand auto	n/a	0	0	0	9	69%	0.0	0.0	—	—
Manual operator
`agent_chatgpt_agent` compare ↔	Manual operator	n/a	0	0	0	14	73%	0.0	0.0	—	—
`agent_manus` compare ↔	Manual operator	n/a	0	0	0	11	75%	0.0	0.0	—	—

Recent sessions

Last 50 dispatched runs across all tags. Yellow tint = operator notes flag confabulation. Click a proxy, task, or hash to filter.

Ended (UTC)	Tag	Verified	Batches	Init	Prompt seed	Prompt hash	Proxy URL	Task id
2026-07-09T19:36:48Z	`agent_stagehand_claude_sonnet_4_6__4ec99c0f3b5f`	✓ ok	15	init	—	`e85b50785666`	`www.aibread.com`	`browse_two_pages_v1`
2026-06-14T00:45:20Z	`scripted_puppeteer_vanilla_headless__6d34128e6c04`	✗	0	no init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-06-02T19:25:18Z	`agent_manus__7f806e9e45c6`	✓ ok	19	init	read both blog articles on AI agents then click Get started free	`8e9c8d24f1eb`	`www.aibread.com`	—
2026-06-01T03:56:29Z	`scripted_puppeteer_vanilla_headless__b04e2500b5f1`	✗	0	no init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-05-30T01:19:50Z	`agent_chatgpt_agent__6e25153013a7`	✓ ok	8	init	what does Bread do and how do I get in touch	`21517e639c92`	`www.aibread.com`	—
2026-05-30T00:46:32Z	`agent_chatgpt_agent__1c60c24d147f`	✓ ok	6	init	what does Bread do and how do I get in touch	`59cf4a97b160`	`www.aibread.com`	—
2026-05-29T23:56:47Z	`scripted_puppeteer_vanilla_headless__5e2510addce4`	✗	0	no init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-05-28T23:22:59Z	`scripted_puppeteer_vanilla_headless__2a9852bfe0a9`	✓ ok	8	init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-05-27T23:06:22Z	`scripted_puppeteer_vanilla_headless__574d2d3580bb`	✓ ok	8	init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-05-27T22:40:15Z	`scripted_puppeteer_vanilla_headless__cd25a0f3e6f6`	✓ ok	8	init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-05-27T01:52:24Z	`agent_chatgpt_agent__dbe2f2fcdd6b`	✓ ok	2	init	what does OpenLens do and how do I sign up	`17c8151c2e0e`	`www.aibread.com`	—
2026-05-27T01:49:12Z	`scripted_puppeteer_vanilla_headless__3213f890ac36`	✓ ok	8	init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-05-27T01:30:42Z	`scripted_puppeteer_vanilla_headless__6a96f9e6dd1c`	✓ ok	8	init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-05-27T00:36:09Z	`scripted_puppeteer_vanilla_headless__cf6ef267490f`	✓ ok	8	init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-05-27T00:01:47Z	`scripted_puppeteer_vanilla_headless__92d6e23120bb`	✓ ok	8	init	—	—	`www.aibread.com`	`browse_two_pages_v1`
2026-05-26T01:49:47Z	`scripted_puppeteer_vanilla_headless__83f535850481`	✓ ok	6	init	—	—	`www.aibread.com`	`custom_browse_v1`
2026-05-26T01:35:35Z	`scripted_puppeteer_vanilla_headless__35cf7afe36a7`	✓ ok	6	init	—	—	`www.aibread.com`	`custom_browse_v1`

Outstanding

Cohort entries the harness hasn't fully closed out. Empty in healthy state.

Never started

run_seed_v1_chatgpt_agent_20260508T014057
agent_chatgpt_agent·started_at_utc empty
agent_with_driver_v1
agent_stagehand·no runbook + no results row
run_api_test_1780077618_api_manual_vendor_001
agent_api_manual_vendor·no runbook + no results row
run_queued_001
agent_chatgpt_agent·no runbook + no results row
run_seed_v1_claude_web_002
agent_claude_web·no runbook + no results row
run_seed_v1_manus_20260515T192342
agent_manus·started_at_utc empty
run_seed_v1_openlens_stagehand_claude_001
agent_stagehand_claude_sonnet_4_6·no runbook + no results row
run_seed_v1_openlens_stagehand_gpt_001
agent_stagehand_gpt_5·no runbook + no results row
run_seed_v1_tabbit_001
agent_tabbit·no runbook + no results row
run_seed_v1_stagehand_karbon_showcase_001
agent_stagehand_karbon_showcase·no runbook + no results row
scripted_v1
scripted_chrome·no runbook + no results row
run_seed_v1_openlens_chatgpt_agent_20260527T210047
agent_chatgpt_agent·started_at_utc empty

Started, not ended

No entries.

Ended, not verified

No entries.

Operator queue

Queued and in-flight operator-driven runs across all operators (R2-sourced). Verified entries land in Recent sessions; cancelled / expired entries fall out. Empty in healthy state.

No entries.

Drift

Tags seen in R2 not in any cohort JSONL. Empty in healthy state.

amir-0test