Applied Lab & Continuous Evolution

Operational Experiments, Agent Workflows & Technical Reviews

Every entry follows a strict editorial unit: the operational question, what was tested, what was actually built or verified, what was learned, where human supervision is required, and how it connects to deployed systems. Proposals and roadmap ideas are explicitly separated from active implementations.

Lab 01 · Agent OperationsStatus: Operational on VPS (Supervised)

Verified: October 2026

Hermes Agent on VPS: Multi-Channel Operational Workflows

Supervised agent execution via Telegram and Discord on dedicated Linux infrastructure

1. Question or Problem

How can an executive operator run persistent, task-specific AI agents accessible from mobile and desktop channels (Telegram and Discord) while maintaining strict infrastructure control and human supervision?

2. What I Tested

Provisioned and configured Hermes Agent (an agent platform, distinct from the underlying LLM APIs it calls) on a dedicated Virtual Private Server (VPS). Connected conversational gateways via Telegram and Discord and tested specialized operational profiles for executive job-market analysis, professional content drafting, English communication refinement, website technical review, and routine task automation.

3. What I Built or Verified

A running VPS-hosted Hermes Agent environment reachable through authenticated Telegram and Discord channels, configured with scoped task profiles that take structured user inputs, execute analysis or web/document inspection tools, and return actionable drafts for human review.

4. What I Learned & Where It Was Applied

Task-specialized profiles with explicit input/output boundaries outperform generic open-ended prompts. Running the agent platform on a dedicated VPS provides full control over logs, environment secrets, and channel access lists.

Application: Used in daily personal operations for job-search qualification, bilingual content preparation (en-US), and technical website reviews.

Operational Boundaries & Human Supervision

Human-in-the-loop supervision is mandatory: the agent prepares research summaries, website audits, or content drafts, but never publishes external communications or submits applications autonomously.

Honest Scope: Proposal vs. Active Implementation

Roadmap / Proposal Stage: A 'Master Orchestrator Profile' to coordinate sub-profiles automatically and a persistent long-term memory layer backed by Obsidian vaults are currently architectural proposals under evaluation—not claimed as completed production features.

Components & Roles in This Lab Entry

Hermes Agent Platform: Core agent runtime and tool execution framework hosted on VPSLinux VPS Infrastructure: Isolated server environment, process supervision, and secret managementTelegram & Discord Gateways: Authenticated mobile and desktop command channelsOpenAI / Gemini Models: Underlying LLM reasoning invoked by Hermes profiles

Connected Case: Enterprise CRM →Connected Case: ProjectFlow PM →

Lab 02 · Security & Digital AuditStatus: Technical Audit & Strategy

Verified: September 2026

Web Security, Digital Visibility & Strategy Review (Templo / Arcade)

Technical assessment of security posture, metadata structure, and organic discoverability

1. Question or Problem

How do technical security posture, metadata hygiene, and content structure intersect when evaluating an existing specialized web property (templo.herniva.es) for digital visibility and safe operation?

2. What I Tested

Conducted a structured review of web architecture, public exposure surface, HTTP/TLS security headers, SEO/metadata signals, and social channel strategy for the Templo arcade project.

3. What I Built or Verified

An actionable technical assessment and digital visibility roadmap covering security hardening recommendations, search indexability improvements, and channel positioning.

4. What I Learned & Where It Was Applied

Even passion or niche web properties benefit from enterprise-style hygiene: clear canonical metadata, hardened response headers, and alignment between web structure and social distribution.

Application: Applied the same metadata, canonical, and security-first review checklist to the redesign of this portfolio and the TimberFlow web engine.

Operational Boundaries & Human Supervision

Documented strictly as an assessment, security review, and strategic proposal rather than a first-party enterprise software deployment.

Components & Roles in This Lab Entry

Web Security & Header Analysis: Review of exposure surface, TLS configuration, and safe defaultsTechnical SEO & Metadata Audit: Evaluation of indexation, OpenGraph tags, and semantic HTML hierarchyDigital & Social Strategy: Audience positioning and channel workflow alignment

Connected Case: TimberFlow Web Engine →Reference Site: templo.herniva.es ↗

Architectural Clarity

Precise Technology Classification & Governance

Senior technology leadership requires distinguishing between agent platforms, development environments, productivity ecosystems, and underlying language models.

Agent Execution Platform (Not an LLM)

Hermes Agent

Hosted on a dedicated Linux VPS and accessed via authenticated Telegram and Discord gateways. Orchestrates task-specific profiles and tool execution under mandatory human-in-the-loop review.

AI-Assisted Development Environment (Not a Runtime Model)

Google Antigravity

Used during software engineering to scaffold, refactor, and verify TypeScript, Next.js, and PostgreSQL implementations while keeping senior human architectural control.

Runtime LLM APIs vs. Conversational Products

OpenAI API vs. ChatGPT / Google Gemini

Distinguishes interactive products (ChatGPT, Gemini workspace) from programmatic model APIs (e.g., OpenAI API wrapper in TimberFlow bounded by rate limits and edge caching).

Enterprise Productivity & Identity Ecosystems

Google Workspace & Microsoft 365

Enterprise collaboration, identity, and document ecosystems—including complex tenant migrations from Google Workspace to Microsoft 365 and automated intake integration.

Want to discuss agent governance, cloud operations, or enterprise AI adoption?

Reach out directly on LinkedIn or explore the deployed project case studies.

Connect on LinkedIn ↗View Deployed Projects