Repository navigation
Releases
CorePulse v0.2.4
CorePulse v0.2.4 — packaging maintenance preview
Update the container's bundled kubectl from v1.35.1 to v1.35.10, the current patch in the same minor line verified from the official Kubernetes release endpoint on 9 November. The lab's Kubernetes server is unchanged. Ordinary Linux CI now builds and executes the image with networking disabled and a read-only root filesystem, including CorePulse's version and kubectl's client-only version.
Native archives now contain the project MIT LICENSE and THIRD_PARTY_NOTICES.md with the Go runtime/standard-library license. The image includes both files under /usr/share/licenses/corepulse/. Separately bundled Kubernetes/base components remain under their upstream licenses; this native notice is not a complete container license inventory. Container publication requests an SBOM attestation. Local untracked Claude workspace metadata is excluded from Docker build contexts; its files are untouched.
CorePulse's Go code and evaluation semantics are unchanged from v0.2.3. That preview introduced prospective session declarations, whole-session holdouts and the read-only doctor command. Original cohort limitations and poor AMF results remain visible; no live model-serving or cross-session accuracy claim is added.
The maintained Go 1.27.3 release toolchain, macOS 13 runtime floor and actual packaged-archive execution gates remain in place. Linux amd64, Windows amd64 and macOS arm64 are natively checked before publication; Linux arm64 and Darwin amd64 remain cross-built without native execution claims.
Changes: https://github.com/lazypanic999/CorePulse/blob/v0.2.4/CHANGELOG.md
Native notices: https://github.com/lazypanic999/CorePulse/blob/v0.2.4/THIRD_PARTY_NOTICES.md
Doctor: https://github.com/lazypanic999/CorePulse/blob/v0.2.4/docs/doctor.md
Session guide: https://github.com/lazypanic999/CorePulse/blob/v0.2.4/docs/session-holdouts.md
This is an isolated-lab preview. Verify archives with SHA256SUMS. Container: ghcr.io/lazypanic999/corepulse:v0.2.4.
Post-publication verification (9 November): four downloaded platform archives matched SHA256SUMS. Windows/Linux archives and the network-isolated, read-only-rootfs container passed both synthetic holdout modes. The exact macOS arm64 archive passed hosted execution; both Darwin archives contain LC_UUID and the macOS 13 minimum. Native project/runtime notices match the repository, kubectl reports v1.35.10, and the registry's SPDX 2.3 attestation contains 92 packages with descriptor digests and image subject verified. Image index digest: sha256:687a78132197a399064b22d62478268fb856caecbc3bd1105332bb3c1566f1b8. The published Linux evaluator preserves the original CSV and every fitted parameter, prediction and metric. Govulncheck v1.9.0 reports no known vulnerabilities in the published CorePulse Windows executable. The live dashboard uses v0.2.4 and its post-reset doctor passed all eight checks.
Detailed verification summary. This is maintainer evidence with explicit scope; raw cohort inputs remain local and the SBOM is not a complete license/security certification.
CorePulse v0.2.3
CorePulse v0.2.3 — session holdouts and lab diagnostics preview
Declare collection occasions prospectively with --session-id in baseline, pod-crash and dashboard startup. Declarations survive interrupted/final checkpoints and optional window metadata. Explicit split --group-by session keeps every member run in one training, held-out or purged group; an embargo purges the entire affected training session. Unknown/insufficient groups, lost class support and forged CSV assignments refuse. Schema-2 saved reports display grouping explicitly; legacy run mode remains schema 1 and its numerical feature allowlist is unchanged.
The synthetic no-cluster example exercises both grouping modes and retains wrong predictions/abstentions. Session names are metadata, not authenticated proof of independent network conditions. The original 22-run study predates session declarations; its missing AMF predictions and stronger binary packet-rule result remain visible. No new cross-session accuracy claim is made.
Add corepulse doctor: eight distinct read-only checks for ownership, context, workloads, packets, N2/N4, current-pod metric coverage and stable pod names. It creates no run, writes no telemetry, performs no repair and never releases a lock. Incomplete observations exit 2; a snapshot cannot authorize bypassing collection preflight. A real tagged baseline/AMF lifecycle test retained the declaration, correctly reported N2 loss despite 2/2 packets, and passed all doctor checks after the ordered reset. Single-session splitting correctly refused its separate exported windows.
Native Windows/macOS CI now accompanies Linux minimum-source-version checks. It exposed a macOS loader failure with the old Go 1.22 linker. Releases and the container use Go 1.27.3; macOS archives require macOS 13 Ventura or newer. Actual packaged Linux amd64, Windows amd64 and macOS arm64 binaries must pass version and both offline walkthrough modes before release/container publication. Linux arm64 and Darwin amd64 are cross-built without native execution checks.
Session guide: https://github.com/lazypanic999/CorePulse/blob/v0.2.3/docs/session-holdouts.md
Doctor: https://github.com/lazypanic999/CorePulse/blob/v0.2.3/docs/doctor.md
No-cluster walkthrough: https://github.com/lazypanic999/CorePulse/blob/v0.2.3/examples/offline/README.md
Measured progress and remaining limits: https://github.com/lazypanic999/CorePulse/blob/v0.2.3/docs/progress/2026-11-09.md
This remains an isolated-lab preview. Packet loss/latency/SCTP injection and live model serving are not implemented. Raw lab telemetry remains local. Verify downloaded archives with SHA256SUMS. Container: ghcr.io/lazypanic999/corepulse:v0.2.3.
Post-publication verification (9 November): SHA256SUMS matched the downloaded Linux/Windows amd64 and both Darwin archives. Windows/Linux packaged binaries ran the complete offline walkthrough in both modes; the Linux evaluator reproduced the original CSV and every fitted parameter, prediction and metric. Both Darwin archives contain LC_UUID and the documented macOS 13 minimum; the exact arm64 archive also passed native hosted execution. The network-disabled, read-only-rootfs container ran the session pipeline against invented inputs only and matched native results. Container digest: sha256:ace9bbb7f40cfe9d55aa0867f36b2368c44bb2882b2c1989f9c53b7c7b4a3b99. Govulncheck v1.9.0 reported no known vulnerabilities in the published CorePulse Windows executable. The running lab dashboard uses this release and its live doctor passed all eight gates after the reset.
CorePulse v0.2.2
Prepare v0.2.2 preview for ownership UI and truthful packet evidence (#53)
Post-publication verification: downloaded Windows/amd64 and Linux/amd64 archives passed SHA256SUMS. The complete synthetic CLI pipeline passed with both published binaries; the Linux binary also observed a healthy 2/2 lab gateway. Rebuilding the frozen real cohort preserved exact window CSV bytes, fitted parameters, predictions and metrics. macOS/ARM archives were built in CI but not executed on native hardware here.
CorePulse v0.2.1
Prepare v0.2.1 preview with measured progress and remaining limits (#49)
Post-publication verification: Windows/amd64 and Linux/amd64 archives were downloaded and checked against SHA256SUMS. The complete synthetic CLI pipeline passed with both published binaries. The Linux binary rebuilt the frozen real window CSV byte-identically and reproduced every fitted parameter, prediction and metric; updated report limitation wording intentionally changes full report bytes. macOS/ARM archives were built in CI but were not executed on native hardware here.
CorePulse v0.2.0
CorePulse v0.1.0
Prepare MIT licensed preview release and GHCR package