AI Baby Gorilla: From Single-Page Identity Diagnostics to Public AI-Agent Execution

This technical report documents the engineering development and validation of AI Baby Gorilla, a single-page JSON-LD identity-connectivity diagnostic developed by Sydney Business Web, and its extension into a publicly discoverable, machine-callable API for AI agents and external software. AI Baby Gorilla examines the structured-data entities, identifiers and relationships on a public webpage, producing an identity-connectivity assessment and an interactive entity graph. Unlike whole-site analysis systems, its diagnostic scope is deliberately limited to one webpage. The investigation documents the transition from a browser-based diagnostic to a Cloudflare Worker API using an Agentic Resource Discovery (ARD) manifest, OpenAPI 3.1 specification and a dedicated HTTP execution endpoint. The engineering implementation includes: Machine-readable discovery: Public ARD manifest and OpenAPI 3.1 interface documentation. Public API execution: Operator-verified anonymous HTTP 200 responses containing structured diagnostic results. Authentication: Optional Bearer credentials, with invalid supplied tokens rejected using HTTP 401. Rate limiting: D1-backed controls enforcing five scans per requester and twenty scans globally per rolling hour. Resource protection: Public URL validation, a 1.5 MB HTML limit, eight-second retrieval timeout and maximum of three redirects. Audit evidence: D1 records of scan admissions, outcomes and enforcement decisions. Agentic Browsing validation: A controlled Lighthouse comparison distinguishing successful ARD discovery auditing from actual API execution. The report also documents controlled WebMCP experimentation and explains the differences between capability discovery, documented invocation and verified execution. Testing demonstrated successful operator-initiated public API calls, authentication rejection, requester rate-limit enforcement and diagnostic response delivery. However, independent autonomous-agent adoption, comprehensive global concurrency testing and externally verified diagnostic accuracy remain unproven. These limitations are explicitly distinguished from implemented capabilities and recorded engineering observations. The investigation establishes a documented, operator-tested architecture for making structured-data identity diagnostics accessible to external software while retaining defined security and resource boundaries. Author: Keith Rowley, Sydney Business Web. Publication: Engineering Research Report 10, Version 1.0.0, 14 November 2026.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-14
DOI
https://doi.org/10.5281/zenodo.24211991
Primary Topic
Semantic Web and Ontologies
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

AI Baby Gorilla: From Single-Page Identity Diagnostics to Public AI-Agent Execution

Keith Rowley
Zenodo (CERN European Organization for Nuclear Research)
Semantic Web and Ontologies
article

AI Baby Gorilla: From Single-Page Identity Diagnostics to Public AI-Agent Execution

Keith Rowley
article en

Abstract

This technical report documents the engineering development and validation of AI Baby Gorilla, a single-page JSON-LD identity-connectivity diagnostic developed by Sydney Business Web, and its extension into a publicly discoverable, machine-callable API for AI agents and external software. AI Baby Gorilla examines the structured-data entities, identifiers and relationships on a public webpage, producing an identity-connectivity assessment and an interactive entity graph. Unlike whole-site analysis systems, its diagnostic scope is deliberately limited to one webpage. The investigation documents the transition from a browser-based diagnostic to a Cloudflare Worker API using an Agentic Resource Discovery (ARD) manifest, OpenAPI 3.1 specification and a dedicated HTTP execution endpoint. The engineering implementation includes: Machine-readable discovery: Public ARD manifest and OpenAPI 3.1 interface documentation. Public API execution: Operator-verified anonymous HTTP 200 responses containing structured diagnostic results. Authentication: Optional Bearer credentials, with invalid supplied tokens rejected using HTTP 401. Rate limiting: D1-backed controls enforcing five scans per requester and twenty scans globally per rolling hour. Resource protection: Public URL validation, a 1.5 MB HTML limit, eight-second retrieval timeout and maximum of three redirects. Audit evidence: D1 records of scan admissions, outcomes and enforcement decisions. Agentic Browsing validation: A controlled Lighthouse comparison distinguishing successful ARD discovery auditing from actual API execution. The report also documents controlled WebMCP experimentation and explains the differences between capability discovery, documented invocation and verified execution. Testing demonstrated successful operator-initiated public API calls, authentication rejection, requester rate-limit enforcement and diagnostic response delivery. However, independent autonomous-agent adoption, comprehensive global concurrency testing and externally verified diagnostic accuracy remain unproven. These limitations are explicitly distinguished from implemented capabilities and recorded engineering observations. The investigation establishes a documented, operator-tested architecture for making structured-data identity diagnostics accessible to external software while retaining defined security and resource boundaries. Author: Keith Rowley, Sydney Business Web. Publication: Engineering Research Report 10, Version 1.0.0, 14 November 2026.

Zenodo (CERN European Organization for Nuclear Research)
Semantic Web and Ontologies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon