Attention First

Attention-first infrastructure tooling. Watch drift. Act now. Remove friction.

Infrastructure tools that surface only what justifies attention. No dashboards. No collection-for-collection's-sake. Quiet when healthy, sharp when not.

Grounded in real production experience.

Watch

Detect systemic risk and trust decay before it becomes an incident.

trustwatch

Kubernetes trust surface monitoring. Discovers expiring certificates on webhooks, API aggregation, and service mesh issuers.

chainwatch

AI agent execution control plane. Enforces safety at irreversible boundaries — payments, credentials, data destruction.

pastewatch

Clipboard obfuscation for macOS. Replaces sensitive data with placeholders before it reaches AI chat interfaces.

Now

Surface only what matters right now. Silence means health.

kubenow

Kubernetes resource analysis and cost optimization. Evidence-based recommendations from Prometheus metrics, not guesswork.

infranow

Real-time infrastructure triage. 13 deterministic detectors, ranked by severity. Empty screen means everything is fine.

Spectre

Compare what code references against what actually exists. Find drift, stale paths, forgotten resources.

vaultspectre

Vault secret path auditor. Scans code for secret references, validates against live Vault, flags stale and missing paths.

clickspectre

ClickHouse usage analyzer. Maps service-to-table relationships, scores cleanup safety, generates interactive reports.

kafkaspectre

Kafka cluster auditor. Finds unused topics, scores cleanup risk, produces batch operation plans.

pgspectre

PostgreSQL schema auditor. Detects unused tables, orphaned indexes, schema drift between code and live database.

mongospectre

MongoDB collection auditor. Finds unused collections, orphaned indexes, missing index coverage.

s3spectre

S3 bucket and object auditor. Detects drift between code references and actual AWS state.

spectrehub

Central aggregator. Collects all spectre tool reports, normalizes issues, scores health, tracks trends.

Contain

Ephemeral incident tooling. Install, capture, uninstall. No permanent footprint.

logtap

Ephemeral log capture and incident triage for Kubernetes. Tap in, grab logs, triage 500GB blind, tap out.

Principle

These tools exist because production systems accumulate silent drift, noisy telemetry, and abandoned artifacts. Instead of collecting everything, they surface only what justifies its cost.

Every tool follows the same rules: deterministic over probabilistic, control over observability, restraint over speed. Tools present evidence and let you decide — mirrors, not oracles.