Marketplace Sync Guard
Local-first, read-only auditing for CSV/XLSX exports. Built to reveal mismatches in SKU, stock, price, presence and status across ERP and marketplace data before they become operational problems.
Open productIndependent software lab Brazil
Lumen Anima Labs is the home for focused software products, automation infrastructure, QA systems and responsible security research built from concrete problems.
Small systems with a clear job instead of generic software for everyone.
Real workflows, measurable behavior and testing before decoration.
Ship, observe, fix the proven bottleneck and keep unnecessary complexity out.
Selected work
Each project below was built around an operational problem and moved beyond a concept. Public products remain available; private systems are shown only at a high level.
Local-first, read-only auditing for CSV/XLSX exports. Built to reveal mismatches in SKU, stock, price, presence and status across ERP and marketplace data before they become operational problems.
Open productA self-contained browser practice lab for evidence-first bug reporting, severity judgment, defensive security awareness, accessibility, responsive testing and truthful portfolio material.
Open QA LabEvidence-first guides that end in a report, checklist, triage decision or test session instead of generic definitions.
Open guidesA price-cliff map for Shopee sellers that shows where charging more can reduce profit and the exact recovery price after each fee boundary.
Open free toolCustom MCP infrastructure connecting ChatGPT to an authorized Windows machine through a private relay and self-healing local agent, with file, process and desktop interaction capabilities.
Core build completed. Now maintained through targeted improvements discovered in real use.
A fast browser word puzzle designed for short playable sessions, with progressive levels, scoring, streaks, responsive interaction and bilingual support.
Open public demoCapabilities
The lab works across product construction and technical validation. The common thread is simple: understand the system, make behavior observable and improve from evidence.
Browser extensions, focused web products and utilities with narrow, useful scopes.
Browser and Windows workflows, process orchestration, MCP tools and custom integrations.
Functional testing, reproducible bug reports, data validation and evidence-driven verification.
Web/API authorization, Broken Access Control, IDOR, role and permission boundaries — only within explicit authorized scope.
AI used as an execution and analysis layer inside practical systems, not as a novelty layer.
Operating method
Start from a workflow, failure mode or repeated cost that can be observed.
Enough engineering to test the premise without weeks of decorative complexity.
Use telemetry, tests, real users or real operating conditions whenever possible.
Fix proven bottlenecks and let working systems stay boring.
About the lab
Lumen Anima Labs is the public home for software products and technical experiments created by Cesar Crispim in Brazil.
The work spans product engineering, automation, QA, applied AI and responsible security research. Projects can be public products, private operational tools or experiments — but they all have to solve a concrete problem before they earn more complexity.