shadownebula

nexus

where science meets the uncharted

systems nominal

// About

shadow-nebula :: classified

$ cat /etc/shadow-nebula/about.txt

Founded at the intersection of technology and the unknown,

Shadow Nebula operates from the dark corners of curiosity.

 

We build, we break, we explore. Our projects span cloud

infrastructure, AI experiments, and things we can't talk

about yet. Some of what you see here is real. Some isn't.

That's by design.

 

The future is forged in darkness and illuminated by insight.

$

// Arsenal

Active projects & experimental systems

ACTIVE SN-001

Lab Portal

Demo & lab environment gateway. Hosts experimental deployments and serves as the proving ground for new infrastructure.

Azure .NET Infrastructure
Access →
RESEARCH SN-002

Nebula Nectar

Perception enhancement protocol. A concoction of machine learning models that see beyond the ordinary noise of raw data.

AI/ML Python Classified
Restricted →
ACTIVE SN-003

Cosmic Codex

Advanced signal interpretation software. Decodes patterns in telemetry, logs, and the ambient hum of distributed systems.

Observability Go Cloud
Coming Soon →
ARCHIVED SN-004

Galactic Grimoire

Wisdom repository inscribed with knowledge from ancient deployments. Infrastructure-as-code templates and battle-tested patterns.

IaC Bicep Terraform
Archived →

// Transmissions

Dispatches from the field

transmission-log :: 2026-03-06

$ cat /var/log/transmissions/006.log

 

DEPLOYING TO THE VOID: AZURE LESSONS LEARNED

 

Three months into running infrastructure on Azure and the

biggest lesson isn't technical — it's about cost discipline.

Every idle resource is a slow leak in the hull.

 

We migrated the Lab Portal from a B1 App Service to a

static-backed deployment. Monthly cost dropped from $13 to

under $1. The trick: treat every service as disposable.

If it can be static, make it static. If it needs compute,

make it serverless. If it needs state, question whether

it really does.

 

Next target: automating the deployment pipeline so pushing

to main triggers a full provision-and-deploy cycle. No more

manual portal clicking.

 

-- END TRANSMISSION --

$

transmission-log :: 2026-02-18

$ cat /var/log/transmissions/005.log

 

THE STARFIELD PROBLEM

 

Spent two days debugging why the starfield canvas was

eating all content on the page. Turns out CSS z-index

doesn't do what you think when you have duplicate

selector blocks overriding each other.

 

The fix was embarrassingly simple: merge the two

.container blocks into one. Position: relative and

z-index: 1 need to live together. The second block

was silently resetting position to static.

 

Lesson: when your content disappears behind a canvas,

check for duplicate CSS selectors before blaming the

browser rendering engine.

 

-- END TRANSMISSION --

$

// Intel

Classified briefings & field reports

INCIDENT REPORT

SUBJECT: Unauthorized Process Injection — CI/CD Pipeline Breach

SITUATION: Automated monitoring flagged anomalous build artifacts in the Sector 9 deployment pipeline at 03:47 UTC. Build hash mismatch detected across ████████ staging environments.

ANALYSIS: Unknown process injected rogue artifacts during the ██████ compilation phase. Signature analysis indicates the payload was designed to establish a persistent ████████████████ callback channel. Origin traced to an IP range associated with ████████████.

CONTAINMENT: Pipeline halted. All artifacts from the affected window purged. Build environment reprovisioned from immutable base image. Secrets rotated across ██ key vaults.

STATUS: Contained. Investigation ongoing. Elevated monitoring in effect for 72 hours.

CLEARANCE: LEVEL 4   DISTRIBUTION: RESTRICTED
INTELLIGENCE BRIEF

SUBJECT: Deep-Space Signal Matches Nebula Nectar Frequency Pattern

SUMMARY: Long-range telemetry array in ████████ detected a repeating signal at 14.7 GHz — an exact match to the harmonic frequency used by the Nebula Nectar perception models for baseline calibration.

SIGNIFICANCE: The Nebula Nectar frequency was generated algorithmically and has never been transmitted externally. The probability of a natural match at this precision is approximately 1 in ██████████. Either someone has reverse-engineered our calibration protocol, or something out there is already tuned to the same wavelength.

ASSESSMENT: Signal origin: unknown. Duration: ██ minutes, repeating at 4-hour intervals. No modulation pattern detected — yet. Nectar team has been notified and is cross-referencing against ████████████ training data.

RECOMMENDATION: Maintain passive monitoring. Do not transmit. Do not acknowledge.

CLEARANCE: LEVEL 3   DISTRIBUTION: NEED-TO-KNOW
INTERCEPTED

SUBJECT: Decoded Transmission — Origin Unknown

SOURCE: ████████████████████████

FREQUENCY: ██████ MHz (non-standard band)

DECODED CONTENT:

THE ██████ IS ALMOST COMPLETE. PHASE TWO BEGINS WHEN THE ████████████ ALIGNS WITH THE OUTER ██████. DO NOT TRUST THE ████████ DATA. REPEAT: THE CALIBRATION ████████████████ IS COMPROMISED. SWITCH TO PROTOCOL ██████ IMMEDIATELY. WE HAVE ████ DAYS.

DECRYPTION CONFIDENCE: 34% — remaining content unrecoverable with current methods.

ANALYST NOTE: Partial keyword matches to ████████ project codenames. Escalated to ████████████ for review.

CLEARANCE: LEVEL 5   DISTRIBUTION: EYES ONLY
PROJECT UPDATE

SUBJECT: Cosmic Codex — Quarterly Status Review

PROJECT: Cosmic Codex (SN-003) — Advanced Signal Interpretation Software

STATUS: Active development. Core telemetry parser at ██% completion. Log anomaly detection module entering ██████ testing phase.

MILESTONES: Pattern recognition engine successfully identified ████ anomalous events in synthetic test data. False positive rate reduced to ██%. Integration with ████████████ observability stack pending security review.

BLOCKERS: Dependency on Nebula Nectar ML models for ████████████████ classification layer. Nectar team reports ██████████ delay due to ████████ incident (see IR-2026-041).

FORECAST: Public release window: ████████████. Internal beta: ██████.

CLEARANCE: LEVEL 2   DISTRIBUTION: INTERNAL
AFTER ACTION

SUBJECT: Visual Cloaking Anomaly — Portal Interface Failure

INCIDENT: Lab Portal interface rendered invisible to all operators for approximately ██ hours. Starfield visualization layer remained functional, creating the appearance that the portal had been replaced by empty space. Multiple operators reported "staring into the void."

ROOT CAUSE: A duplicate CSS selector block caused the rendering context to silently reset. The interface content was physically present but existed at a lower z-plane than the visualization layer — effectively cloaked. The ████████████ rendering pipeline did not flag the conflict because both blocks individually passed validation.

RESOLUTION: Merged conflicting selector blocks. Restored interface visibility. Added monitoring for ████████ z-plane conflicts to prevent recurrence.

LESSONS LEARNED: When an interface goes dark, check the stacking context before assuming ████████████████ hostile action. Sometimes the simplest bugs are the most effective camouflage.

CLEARANCE: LEVEL 2   DISTRIBUTION: ALL HANDS

// Project File

Classified project documentation

← Back to Arsenal
RESTRICTED
SN-002 CLASSIFICATION: LEVEL 4

PROJECT FILE: Nebula Nectar — Perception Enhancement Protocol

OVERVIEW: Nebula Nectar is a suite of machine learning models designed to detect patterns invisible to conventional analysis. Originally developed for ████████████████ signal processing, the system has evolved into a general-purpose perception enhancement framework.

CAPABILITIES: The current iteration operates across ██ data modalities including telemetry streams, image analysis, and ████████████ natural language corpus mining. Detection confidence exceeds ██% on benchmark datasets.

ARCHITECTURE: Built on a custom ████████████████ transformer architecture with ████ attention heads. Training infrastructure runs on ██████████ GPU cluster provisioned through ████████ Azure regions. Total training cost to date: ████████████.

CURRENT STATUS: Active research. Version ██████ deployed to internal sandbox. External access denied pending security review of ████████████████████████ output filtering mechanisms.

KNOWN RISKS: Model exhibits ████████████████ behavior when processing data from ████████ sources. Research team investigating whether this constitutes genuine anomaly detection or ████████████ hallucination.

ACCESS: Restricted to Level 4+ clearance. All interactions logged. No external API access authorized.

// Project File

Classified project documentation

← Back to Arsenal
IN DEVELOPMENT
SN-003 CLASSIFICATION: LEVEL 2

PROJECT FILE: Cosmic Codex — Advanced Signal Interpretation Software

OVERVIEW: Cosmic Codex decodes patterns in telemetry, application logs, and distributed system metrics. Where traditional observability tools show you what happened, Codex aims to show you what's about to happen.

APPROACH: Built in Go for performance. Ingests structured and unstructured log streams, applies ████████████ pattern matching, and surfaces anomalies through a custom scoring engine. Integration targets include Prometheus, ████████, and Azure Monitor.

PROGRESS: Core telemetry parser: ██% complete. Anomaly detection module: entering beta testing. Signal classification layer blocked pending Nebula Nectar ████████ model delivery.

DEPENDENCIES: Requires Nebula Nectar v██████ for ML-assisted classification. Current workaround uses rule-based heuristics with ██% accuracy gap.

TIMELINE: Internal beta: ████████████. Public release: pending. The team is ████████████████ optimistic but ████████ realistic.

ACCESS: Development access granted to Level 2+. Source code repository: ████████████████████████.

// Project File

Classified project documentation

← Back to Arsenal
ARCHIVED
SN-004 CLASSIFICATION: LEVEL 1

PROJECT FILE: Galactic Grimoire — Infrastructure Wisdom Repository

OVERVIEW: The Galactic Grimoire was a curated collection of infrastructure-as-code templates, deployment patterns, and operational runbooks distilled from ████ production incidents and ████████ deployments.

CONTENTS: ██ Bicep modules, ██ Terraform configurations, ████ pipeline templates, and ██ disaster recovery playbooks. Covered Azure, ████████, and hybrid environments.

WHY ARCHIVED: The rapid evolution of IaC tooling rendered many patterns obsolete. Rather than maintain increasingly stale documentation, the decision was made to archive the repository and extract still-relevant patterns into active project codebases.

LEGACY: Key patterns from the Grimoire live on in the Lab Portal deployment pipeline and the ████████████ provisioning system. The repository remains read-accessible for historical reference.

ACCESS: Read-only. No further updates planned. Repository: ████████████████████████.

// Transmissions cards

Dispatches from the field

06 MAR 2026

Deploying to the Void: Azure Lessons Learned

Azure Infrastructure Cost

Three months into running infrastructure on Azure and the biggest lesson isn't technical — it's about cost discipline. Every idle resource is a slow leak in the hull. We migrated the Lab Portal from a B1 App Service to a static-backed deployment and monthly cost dropped from $13 to under $1.

Read transmission →
18 FEB 2026

The Starfield Problem

CSS Canvas Debugging

Spent two days debugging why the starfield canvas was eating all content on the page. Turns out CSS z-index doesn't do what you think when you have duplicate selector blocks overriding each other. The fix was embarrassingly simple.

Read transmission →

// Transmissions Dossier

Dispatches from the field

DECLASSIFIED

SUBJECT: Deploying to the Void — Azure Lessons Learned

SITUATION: Three months into running infrastructure on Azure. Cost overruns detected on non-critical services. Review initiated.

FINDINGS: Lab Portal running on B1 App Service Plan — monthly cost $13.14. Service serves only static content. No compute required.

ACTION TAKEN: Migrated to static-backed deployment. Monthly cost reduced to ████ $0.87. Applied same pattern to ██████████ additional services.

RECOMMENDATION: Treat every service as disposable. If it can be static, make it static. If it needs compute, make it serverless. If it needs state, ████████ question whether it really does.

CLEARANCE: LEVEL 2   DISTRIBUTION: INTERNAL
RESTRICTED

SUBJECT: The Starfield Problem — CSS Z-Index Anomaly

SITUATION: Starfield canvas rendering correctly but all page content invisible. Users reported blank screen with only particle animation visible.

ANALYSIS: Root cause identified as duplicate .container CSS selector blocks. Second block silently reset position to static, nullifying z-index stacking context. Content rendered behind canvas at z-index 0.

RESOLUTION: Merged duplicate selector blocks. Confirmed position: relative and z-index: 1 applied correctly. Content restored. ████████████ no further incidents reported.

LESSON: When content disappears behind a canvas element, check for ██████ duplicate CSS selectors before investigating browser rendering engine bugs.

CLEARANCE: LEVEL 3   DISTRIBUTION: NEED-TO-KNOW