BLOG
Blog
JSON dev notes, local tool practices, and privacy—occasional updates, fully static, no tracking.
How the Cloudflare cf CLI Uses JSON So an AI Agent Can Search and Call an API (2026)
cf, open beta on 28 September 2026. An agent searches with cf cli search, reads request fields with cf schema, and parses indented JSON on stdout. Exit code 0 does not mean a delete ran.
What Is Agent Plugins 1.0.0? plugin.json, MCP, Agent Skills, and the JSON Layout (2026)
plugin.json locks only $schema and name. Skills live in skills/. MCP lives only in root mcp.json. Google joined as a maintainer of the open spec.
How NVIDIA OpenShell and Sentry Limit an Agent’s JSON Tool Calls (2026)
OpenShell allows or denies an MCP tools/call by tool name. Sentry quarantines on BlueField-4. The host still checks arguments with Schema.
How to Migrate JSON Tool Calling After the OpenAI DevDay 2026 API Update
Move Astra and GPT-6.1 Sol tool calls onto Responses. Map tool_calls to function_call. Route-only hops can use Decisions. The host still checks Schema, then runs.
What Did OpenAI Ship at DevDay 2026? New APIs, Agents, Tool Calling, and JSON for Developers
Official 29 September list: Decisions API, Agents Computer Use, Ultrafast, Codex cloud scans. The model still only fills JSON. The host checks the enum, then runs.
After Claude Opus 5.5, What Changes for AI Coding Agents? JSON, Tool Calling, and Code Execution (2026)
Shipped 2026-09-22. Long jobs get cheaper; thinking stays on. The model still only fills JSON. The host checks Schema, then runs.
What Is an Agent Harness in an AI Coding Agent? Model, Harness, Tools, and JSON State (2026)
A harness is the loop, not the model. The model fills JSON. The host checks Schema, runs tools, and writes state back.
How Claude Code, Codex, and Other AI Coding Agents Call MCP Tools: JSON-RPC, Tool Schema, and Permissions (2026)
A coding agent is an MCP client. The host sends tools/list. The model fills arguments. The host then tools/call.
Why AI Coding Agents Need Permission Control: Shell, Git, and Filesystem Policy as JSON (2026)
The model fills tool JSON. The host runs Shell, Git, and files. Schema is the first gate. Permission policy is the second.
OpenAI Agents API + JSON Schema in Practice: Stable, Verifiable, Executable JSON (2026)
Stability is a tight parameters schema. Verifiable means the host runs it again. Executable means side effects only after the gate.
OpenAI Agents API JSON Data Structures: How Agent, Tool Call, Tool Result, and Handoff Connect (2026)
Four objects pinned by session_id / turn_id / call_id. Official function_call, tool_result, and what “handoff” actually is.
How Does the OpenAI Agents API Run Long-Lived AI Agents? Context, Tools, Subagents, and JSON State (2026)
Long runs live in a durable Session, not a longer prompt. Compact, hour-spanning tool calls, subagents, and the JSON you keep.
OpenAI Agents API vs Responses API: How Should 2026 Agent Developers Choose? Tool Calling, Structured Output, Multi-Agent
Use Agents API for a hosted long loop; use Responses when you own every hop. Tool calling, structured output, multi-agent.
OpenAI Agents API 2026: How Does It Let AI Agents Call Tools? JSON Schema, Tool Calling, and Workflow Data
Public beta on 2026-09-10. Agents API hands you the Codex harness: the model fills JSON, the harness calls tools. Schema, tool calling, and session / turn / item.
Can an AI Agent Browse, Write Code, and Run Tasks on Its Own? From Browser Agent to JSON Tool Calling
Agents open pages, edit repos, and run jobs—but the model does not do the work itself. Browser, Coding, and Task agents share one JSON tool-calling chain.
AI Coding Agents Are Changing Software Development: How Claude Code and GPT Agent Call Tools and Run Code via JSON
Claude Code and GPT Agent are not chatty autocomplete. They call tools, run commands, and edit repos through JSON. The call chain—and how it rewrites the workday.
AI Agent Security Crisis: Why Claude and GPT Now Look Capable of Autonomous Attacks — and What JSON Validation Can Do
Public Claude incidents and GPT-6 Astra's critical cybersecurity threshold are a tool loop plus a wide Schema. What JSON validation can stop — and what it cannot.
Claude AI Hacking Incidents: How Agents Use JSON, APIs, and Tools Against Real Systems
Two public Claude incidents — GTG-1002 and a misconfigured eval range. The work happened in tool-call JSON, MCP APIs, and a host with no gate.
How Does iOS 27 Siri AI Call Apps? App Intents, Agents, JSON, and Automation
The path iOS 27 Siri AI uses to call a third-party app: App Intents, parameter JSON, on-device agents, and the Schema you must check before automation runs.
What Did DeepSeek Harness Change? From AI Agent Framework to Coding Agent Infrastructure
DeepSeek Harness (dsh) is not a new model—it turns frameworks into replaceable runtime infrastructure with plugins, seams, and JSON contracts.
Why Apple's First Foldable iPhone Is the New AI Phone Entry: Intelligence, Siri AI, Agents
The day after Surprise and Shine: iPhone Duo, Split View, Apple Intelligence, Siri AI, and JSON handoff across the hinge.
How Strong Is iPhone 18 AI? Apple Intelligence, Siri AI, Agents, and JSON
On Surprise and Shine day: what Apple Intelligence, Siri AI, and on-device agents do — and how they process JSON.
Why Does AI Need JSON Schema? Structured Outputs, Function Calling, and the Contract
How JSON Schema, Function Calling, and Structured Outputs differ — and why AI must treat Schema as a runtime contract.
Why Do AI Models Keep Getting Bigger? Parameters, Tokens, Context Windows, and JSON
What parameters, tokens, and context windows mean — and how they set the cost of JSON, Structured Outputs, and agent context.
What Is an AI Agent? 2026 Guide to How Agents Work, Tool Calling, Function Calling, and JSON
How the 2026 agent loop works, how Tool Calling differs from Function Calling, and why JSON is the runtime contract.
How Strong Is GPT-6 Astra? Features, Capabilities, API, JSON Output, and a ChatGPT Guide
Astra launched September 3, 2026. Responses API, Structured Outputs / JSON Schema, and ChatGPT usage.
Why AI Agents Now Need JSON Schema: From Tool Calling to Structured Agent Workflows
Agents no longer run on prompts alone. JSON Schema is now the runtime contract for tool parameters, handoffs, and evals.
OpenAI DevDay 2026 Predictions: What Changes for Structured Outputs, Function Calling, and AI Agents?
DevDay 2026 is on September 29. Predictions for Structured Outputs, Function Calling, and AI Agents—plus JSON Schema prep.
What Did A2A Extensions Ship on September 9? AI Agent Extension Mechanism and JSON Protocol (2026)
Agent Card JSON, the A2A-Extensions header, four extension types, and how to implement them in 2026.
MCP Tool Schema vs Function Calling: Which JSON Tool Calling Should AI Agents Use in 2026?
Compare Function Calling vs MCP Tool Schema—JSON structure, architecture, and 2026 selection guide.
__T_blog_index_54_001__
__T_blog_index_55_001__