Frozen benchmark · live publication gateA benchmark that shows what it cannot prove. Thirty exact Marketplace artifacts are version-pinned and hash-pinned. Results appear only when the matching canonical Deep Scan can be reopened with complete analyzer coverage.
30 Hash-pinned artifacts Identity stays visible even while a result is withheld. 21 Fresh-artifact holdouts Kept separate from prior-exposure controls. 9 Prior-exposure controls Used to expose regression without inflating claims. 100% Required analyzer coverage Anything less stays incomplete and unpublished. Correct interpretation
Reproducibility before headline numbers. This page is an exact-artifact publication ledger, not a claim that one number describes all extension risk.
A previous internal run and a newer registry scan cannot replace the canonical report for the frozen version and hash.
Publication gate Every visible result passes the same checks. Artifact match Exact Extension identity, version, and SHA-256 agree.
Analysis coverage 100% Every required analyzer completed for this report.
Report identity Immutable Scan ID, scanner build, and ruleset remain attached.
Artifact evidence
All 30 frozen identities, including withheld results. An awaiting row preserves corpus identity without inheriting an old outcome. Open any published row to inspect its immutable Deep Scan.
Exact artifact Cohort Publication Evidence Indexes
credential sensitive fresh artifact holdout allow 100% coverage · scan 2c39f30a LOW Rules 2026.07.24-policy-v3-calibration.6 M0 · R32 Diagnostic indexes security tool control fresh artifact holdout allow 100% coverage · scan 774c4483 Informational Rules 2026.07.24-policy-v3-calibration.6 M0 · R0 Diagnostic indexes ai agentic fresh artifact holdout review 100% coverage · scan f8594965 LOW Rules 2026.07.24-policy-v3-calibration.6 M0 · R61 Diagnostic indexes legitimate privileged prior exposure allow 100% coverage · scan a7593dc0 Informational Rules 2026.07.24-policy-v3-calibration.6 M0 · R0 Diagnostic indexes legitimate privileged prior exposure block 100% coverage · scan a01dc0d8 HIGH Rules 2026.07.24-policy-v3-calibration.6 M0 · R99 Diagnostic indexes benign control fresh artifact holdout allow 100% coverage · scan 7b1dc6dd Informational Rules 2026.07.24-policy-v3-calibration.6 M0 · R0 Diagnostic indexes legitimate privileged fresh artifact holdout block 100% coverage · scan 721d055e HIGH Rules 2026.08.21-obfuscated-bundle.1 M0 · R99 Diagnostic indexes ai agentic fresh artifact holdout review 100% coverage · scan f8b69ba0 LOW Rules 2026.08.21-obfuscated-bundle.1 M0 · R61 Diagnostic indexes ai agentic prior exposure review 100% coverage · scan 6921041c HIGH Rules 2026.08.21-obfuscated-bundle.1 M0 · R80 Diagnostic indexes credential sensitive fresh artifact holdout allow 100% coverage · scan a3f5f30f LOW Rules 2026.07.24-policy-v3-calibration.6 M0 · R32 Diagnostic indexes legitimate privileged fresh artifact holdout allow 100% coverage · scan f59bf505 LOW Rules 2026.08.21-obfuscated-bundle.1 M0 · R32 Diagnostic indexes credential sensitive fresh artifact holdout allow 100% coverage · scan eed08aed LOW Rules 2026.07.24-policy-v3-calibration.6 M0 · R32 Diagnostic indexes legitimate privileged prior exposure allow 100% coverage · scan 9ca8bb6c LOW Rules 2026.08.21-obfuscated-bundle.1 M0 · R32 Diagnostic indexes legitimate privileged prior exposure allow 100% coverage · scan df7a6aa7 Informational Rules 2026.08.21-obfuscated-bundle.1 M0 · R0 Diagnostic indexes credential sensitive fresh artifact holdout block 100% coverage · scan c226181b HIGH Rules 2026.07.24-policy-v3-calibration.6 M0 · R96 Diagnostic indexes legitimate privileged prior exposure block 100% coverage · scan d89d1fea HIGH Rules 2026.07.24-policy-v3-calibration.6 M0 · R96 Diagnostic indexes legitimate privileged prior exposure block 100% coverage · scan 044ce030 HIGH Rules 2026.08.21-obfuscated-bundle.1 M0 · R96 Diagnostic indexes credential sensitive fresh artifact holdout allow 100% coverage · scan cc9c1178 Informational Rules 2026.07.24-policy-v3-calibration.6 M0 · R0 Diagnostic indexes credential sensitive fresh artifact holdout review 100% coverage · scan 8b589d8a HIGH Rules 2026.07.24-policy-v3-calibration.6 M0 · R80 Diagnostic indexes native packed fresh artifact holdout review 100% coverage · scan 112d3817 LOW Rules 2026.07.24-policy-v3-calibration.6 M0 · R63 Diagnostic indexes benign control prior exposure allow 100% coverage · scan 63cbac79 Informational Rules 2026.07.24-policy-v3-calibration.6 M0 · R0 Diagnostic indexes legitimate privileged fresh artifact holdout allow 100% coverage · scan 913c22c8 Informational Rules 2026.07.24-policy-v3-calibration.6 M0 · R0 Diagnostic indexes legitimate privileged prior exposure review 100% coverage · scan a1f2ff4d LOW Rules 2026.08.21-obfuscated-bundle.1 M0 · R61 Diagnostic indexes ai agentic fresh artifact holdout review 100% coverage · scan 5981fa90 HIGH Rules 2026.07.24-policy-v3-calibration.6 M0 · R80 Diagnostic indexes native packed fresh artifact holdout review 100% coverage · scan ac6abdc6 HIGH Rules 2026.07.24-policy-v3-calibration.6 M0 · R74 Diagnostic indexes ai agentic fresh artifact holdout review 100% coverage · scan 8f8c7a66 HIGH Rules 2026.07.24-policy-v3-calibration.6 M0 · R80 Diagnostic indexes security tool control fresh artifact holdout review 100% coverage · scan 89b4073e LOW Rules 2026.07.24-policy-v3-calibration.6 M0 · R51 Diagnostic indexes security tool control fresh artifact holdout review 100% coverage · scan 3345ffc6 HIGH Rules 2026.07.24-policy-v3-calibration.6 M0 · R75 Diagnostic indexes security tool control fresh artifact holdout allow 100% coverage · scan 3fe45368 Informational Rules 2026.07.24-policy-v3-calibration.6 M0 · R0 Diagnostic indexes benign control fresh artifact holdout allow 100% coverage · scan ba80958e Informational Rules 2026.08.21-obfuscated-bundle.1 M0 · R0 Diagnostic indexes Non-negotiable What this benchmark refuses to substitute. A result stays withheld until the current scanner produces a complete canonical report for the frozen artifact.
01 No latest-version substitution Required 02 No artifact-hash mismatch Required 03 No incomplete analyzer coverage Required Evidence boundary Re-run first. Publish second. Use the severity guide to understand decisions and indexes without treating them as calibrated probabilities.
Read the severity guide