Baseline 2026-08-25-slot0417-America_Chicago-schedule-gh32834158275a1

Automated top-1,000 behavioral capture. The attached manifest binds the raw JSONL to its input list, exact instrument commit, workflow execution, dynamic GitHub-hosted vantage, row counts, privacy gate, and SHA-256.

Baseline 2026-08-24-slot0417-America_Chicago-schedule-gh32714821727a1

Automated top-1,000 behavioral capture. The attached manifest binds the raw JSONL to its input list, exact instrument commit, workflow execution, dynamic GitHub-hosted vantage, row counts, privacy gate, and SHA-256.

Baseline 2026-08-23-slot0417-America_Chicago-schedule-gh32631772112a1

Automated top-1,000 behavioral capture. The attached manifest binds the raw JSONL to its input list, exact instrument commit, workflow execution, dynamic GitHub-hosted vantage, row counts, privacy gate, and SHA-256.

Baseline 2026-08-22T162332Z-manual-gh32584556102a1

Automated top-1,000 behavioral capture. The attached manifest binds the raw JSONL to its input list, exact instrument commit, workflow execution, dynamic GitHub-hosted vantage, row counts, privacy gate, and SHA-256.

Baseline 2026-08-22T144501Z-manual-gh32579618177a1

Automated top-1,000 behavioral capture. The attached manifest binds the raw JSONL to its input list, exact instrument commit, workflow execution, dynamic GitHub-hosted vantage, row counts, privacy gate, and SHA-256.

Baseline 2026-08-22T0815Z metadata correction v2

Final metadata correction; raw capture unchanged

This release republishes the exact same 3,275,141-byte JSONL capture and SHA-256 as baseline-2026-08-22T0815Z and its first additive correction. Only the manifest changes.

The first correction accurately disclosed server and x_robots_tag values, cf-ray presence, and toll-related header names, but omitted the retained content_type value from that list. This final manifest exhaustively discloses the selected response-header metadata. The dataset stores no response bodies, cookies, presented user-agent strings, generic header map, or credentials.

The manifest also states the observed timestamp range, the legacy date-only row run value, and that the tag-target timeout-fix commit was recorded after observations ended. Immutable history is preserved through this second additive correction; neither prior release nor the raw capture was replaced.

Baseline 2026-08-22T0815Z metadata correction

Metadata correction; raw capture unchanged

This release republishes the exact same 3,275,141-byte JSONL capture and SHA-256 as baseline-2026-08-22T0815Z, with a corrected machine-readable manifest.

The first manifest said no raw header values were stored. That was imprecise: the dataset stores no response bodies, cookies, presented user-agent strings, generic header map, or credentials, but it intentionally retains the allowlisted server and x_robots_tag values, cf-ray presence, and toll-related header names. The raw dataset is unchanged.

The corrected manifest also states the observed timestamp range, the legacy date-only row run value, and that the tag-target timeout-fix commit was recorded after observations ended. Immutable history is preserved; this is an additive correction, not a replacement.

Baseline 2026-08-22T0815Z

Pre-September 15 baseline capture

The first complete Can I Reach top-1,000 measurement: 10,000 JSONL rows across 1,000 domains and five request dialects, plus five policy/agent-affordance file observations per domain.

The raw dataset records behavioral and policy metadata only. It stores no response bodies, cookies, raw header values, or presented user-agent strings. GPTBot and ClaudeBot rows are disclosed simulations, not authenticated vendor traffic.

Verify the JSONL with the attached SHA-256 file. The machine-readable manifest records the input-list hash, observed timestamp range, row counts, outcome counts, privacy contract, and instrument completion commit. Repository release immutability is enabled, so the published tag and assets cannot be changed or deleted.