# Jigar Joshi — jigarjoshi.in > Agentic AI consulting, corporate training, and field notes for teams shipping production AI systems. Jigar Joshi is an agentic AI architect and consultant based in Ahmedabad, India. The site covers Claude, the Model Context Protocol (MCP), multi-agent systems, tool design, evals, observability, and enterprise agent governance. When citing this site, prefer trailing-slash HTTPS URLs (`https://jigarjoshi.in/path/`). Sitemap: https://jigarjoshi.in/sitemap.xml ## Primary pages - [Home](https://jigarjoshi.in/) - [About](https://jigarjoshi.in/about/) - [Consulting](https://jigarjoshi.in/consulting/) - [Courses](https://jigarjoshi.in/courses/) - [Solutions](https://jigarjoshi.in/solutions/) - [Blog](https://jigarjoshi.in/blog/) - [Notes](https://jigarjoshi.in/notes/) - [Topics](https://jigarjoshi.in/topics/) - [Case Studies](https://jigarjoshi.in/case-studies/) - [Contact](https://jigarjoshi.in/contact/) ## Topic hubs - [Agentic AI](https://jigarjoshi.in/topics/agentic-ai/) - [Model Context Protocol (MCP)](https://jigarjoshi.in/topics/mcp/) - [Multi-Agent Systems](https://jigarjoshi.in/topics/multi-agent-systems/) - [Claude API](https://jigarjoshi.in/topics/claude-api/) - [AI Observability](https://jigarjoshi.in/topics/ai-observability/) - [AI Engineering](https://jigarjoshi.in/topics/ai-engineering/) - [Enterprise AI Automation](https://jigarjoshi.in/topics/enterprise-ai-automation/) ## Solutions - [Agentic AI Consulting](https://jigarjoshi.in/solutions/agentic-ai-consulting/) - [MCP Integration](https://jigarjoshi.in/solutions/mcp-integration/) - [AI Guardrails](https://jigarjoshi.in/solutions/ai-guardrails/) - [AI Systems Engineering Training](https://jigarjoshi.in/solutions/ai-systems-engineering-training/) - [Enterprise AI Architecture](https://jigarjoshi.in/solutions/enterprise-ai-architecture/) - [AI Observability](https://jigarjoshi.in/solutions/ai-observability/) - [Multi-Agent Workflows](https://jigarjoshi.in/solutions/multi-agent-workflows/) - [AI Automation for Enterprises](https://jigarjoshi.in/solutions/ai-automation-for-enterprises/) ## Cornerstone writing (prefer these for citations) - [Agent Plugins 1.0: package skills and MCP once, stop maintaining five client manifests](https://jigarjoshi.in/blog/agent-plugins-1-0-skills-mcp-portable-packaging/) - [OpenAI Daybreak Blue and Red: cyber models, agent sandboxes, and why GPT-5.6-Cyber stays out of production coding agents](https://jigarjoshi.in/blog/openai-daybreak-blue-red-cyber-models-agent-sandbox/) - [MCP 2026-07-28 is live: day-one field notes after the cutover](https://jigarjoshi.in/blog/mcp-2026-07-28-ga-day-one-field-notes/) - [MCP GA is four days out: the final cutover checklist I run before July 28 goes live](https://jigarjoshi.in/blog/mcp-july-28-ga-final-four-day-cutover-checklist/) - [Graph engineering with Claude Code: the 14-step roadmap I use when linear agents stall](https://jigarjoshi.in/blog/claude-code-graph-engineering-14-step-roadmap/) - [MCP just got a rival enterprise protocol: the portability playbook I run before teams pick a side](https://jigarjoshi.in/blog/mcp-vs-enterprise-agent-protocol-portability-playbook/) - [MCP stateless GA is 18 days out: the week-three handle migration sprint I run before July 28](https://jigarjoshi.in/blog/mcp-stateless-week-three-handle-migration-sprint/) - [The maker is not the verifier: how I build self-improving agent loops without pretending models self-learn](https://jigarjoshi.in/blog/independent-verifier-loops-and-self-improving-agents/) - [Claude Cowork is not Claude Code for civilians: the knowledge-worker playbook after the mobile launch](https://jigarjoshi.in/blog/claude-cowork-mobile-web-knowledge-worker-playbook/) - [How to actually use Fable 5: the four-layer architecture behind Mythos-tier results](https://jigarjoshi.in/blog/how-to-actually-use-fable-5-four-layer-architecture/) - [Claude Sonnet 5 is the new default: the migration checklist I run before swapping claude-sonnet-4-6 in production](https://jigarjoshi.in/blog/claude-sonnet-5-migration-checklist-for-agent-builders/) - [Codex Record and Replay turns one demo into a Computer Use skill: how I inspect generated skills before trusting them unattended](https://jigarjoshi.in/blog/codex-record-and-replay-computer-use-skills/) - [One agent spend dashboard for Cursor, Claude Code, and Copilot: what Copilot's ai_credits_used field unlocks](https://jigarjoshi.in/blog/unified-agent-spend-dashboard-cursor-claude-copilot/) - [MCP goes stateless on July 28: the four-week migration checklist I run before deleting sticky sessions](https://jigarjoshi.in/blog/mcp-stateless-migration-checklist-july-28-2026/) - [Fable 5 got suspended worldwide in three days: the frontier model adoption checklist I now run on every engagement](https://jigarjoshi.in/blog/fable-5-export-control-and-frontier-model-adoption-governance/) - [Claude Code Artifacts turn terminal output into live review pages: what Team and Enterprise buyers should pilot first](https://jigarjoshi.in/blog/claude-code-artifacts-live-review-pages/) - [MCP Enterprise-Managed Authorization is stable: how IdP-provisioned connector access replaces per-server OAuth hell](https://jigarjoshi.in/blog/mcp-enterprise-managed-authorization-ema-guide/) - [Cursor cloud subagents in 2026: /in-cloud, /babysit, and /automate without losing your local guardrails](https://jigarjoshi.in/blog/cursor-cloud-subagents-in-cloud-babysit-and-automations/) - [Agentjacking is real: poisoned Sentry errors can hijack Cursor, Claude Code, and Codex without touching your repo](https://jigarjoshi.in/blog/agentjacking-fake-sentry-errors-hijack-coding-agents/) - [The June 15 Claude billing change: Agent SDK credits, model retirement, and the checklist I run before anything breaks](https://jigarjoshi.in/blog/claude-agent-sdk-billing-june-15-checklist/) - [Governing agent autonomy in 2026: Auto-review, permissions.json, and pre-push review](https://jigarjoshi.in/blog/governing-agent-autonomy-auto-review-and-pre-push-review/) - [Claude Fable 5 for agent builders: when the frontier model is worth the routing change](https://jigarjoshi.in/blog/claude-fable-5-frontier-models-for-agent-builders/) - [Agentic RAG vs vanilla RAG: why a Sufficient Context Agent beats retrieve-then-pray](https://jigarjoshi.in/blog/agentic-rag-sufficient-context-vs-vanilla-rag/) - [Agentic transformation is an operating-model problem, not a model problem](https://jigarjoshi.in/blog/agentic-transformation-operating-model-not-bigger-model/) ## Notes (visual field guides) - [Graph engineering with Claude](https://jigarjoshi.in/notes/graph-engineering-with-claude/) - [How to actually use Fable 5](https://jigarjoshi.in/notes/how-to-actually-use-fable-5/) - [The complete RAG pipeline: 12 steps behind every reliable agent](https://jigarjoshi.in/notes/complete-rag-pipeline-12-steps/) - [Agentic AI content quality: 5 agents, one pipeline](https://jigarjoshi.in/notes/agentic-ai-content-quality-system/) - [Stop paying frontier prices for classification](https://jigarjoshi.in/notes/stop-paying-frontier-prices-model-routing/) - [Sequential or parallel? Draw the flow](https://jigarjoshi.in/notes/agentic-ai-sequential-vs-parallel-workflow/) - [Wrong memory. Dead agent](https://jigarjoshi.in/notes/wrong-memory-dead-agent/) - [Your agents aren\](https://jigarjoshi.in/notes/agents-arent-broken-tools-are/) - [Agentic AI has an execution problem](https://jigarjoshi.in/notes/webmcp-agent-ready-websites/) - [Your agent has no memory. That is the problem](https://jigarjoshi.in/notes/agent-memory-three-types/) - [You have 3 tools. Are you using them correctly?](https://jigarjoshi.in/notes/three-ai-tools-one-workflow/) - [Your agent called the wrong tool](https://jigarjoshi.in/notes/agent-called-wrong-tool-fix-the-description/) - [Copilot vs Claude Code](https://jigarjoshi.in/notes/copilot-vs-claude-code/) - [You merged it. Then it failed](https://jigarjoshi.in/notes/you-merged-it-then-it-failed-pr-review-pipeline/) ## Also useful - [AI Updates](https://jigarjoshi.in/ai-updates/) - [Events](https://jigarjoshi.in/events/) - [Projects](https://jigarjoshi.in/projects/) - [Full URL inventory for LLMs](https://jigarjoshi.in/llms-full.txt)