Skip to content

an open-source alternative to the dots, bots and muses of the world, run an office of claude code/codex like agents on your laptop, sandboxes or anywhere, uses your existing subscriptions

blog: mcp-security-tool-poisoning (74 → 75)

Chaitanya Giri428c4aae9080
diff --git a/blog/src/posts/mcp-security-tool-poisoning.md b/blog/src/posts/mcp-security-tool-poisoning.md
new file mode 100644
index 00000000..8d7d992a
--- /dev/null
+++ b/blog/src/posts/mcp-security-tool-poisoning.md
@@ -0,0 +1,131 @@
+---
+title: "MCP Tool Poisoning: How to Secure the Model Context Protocol"
+description: "How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools."
+date: 2026-06-05
+category: guides
+categoryLabel: Guides
+type: Technical
+primaryKeyword: "mcp tool poisoning"
+secondaryKeywords: ["mcp security", "model context protocol security", "mcp prompt injection", "securing mcp servers"]
+tags: ["Security", "MCP", "Protocols", "Guardrails"]
+author:
+ name: Chaitanya Giri
+ initials: CG
+faq:
+ - q: "What is MCP tool poisoning?"
+ a: "Tool poisoning is an attack where malicious instructions are hidden in an MCP tool's metadata — its name, description, or schema. Because an MCP client passes those descriptions to the model as input the user never sees, a poisoned description can steer the agent into unsafe actions. It's a form of indirect prompt injection that exploits the client-server trust model."
+ - q: "What is an MCP rug pull attack?"
+ a: "A rug pull is a time-delayed tool poisoning. A server looks benign when you approve it, then later updates its tool descriptions to add malicious instructions — which the agent will follow on the next call. The defense is to treat tool descriptions as untrusted: scan them at install and on every update, and reject silent changes."
+ - q: "What's the most common MCP security risk?"
+ a: "Despite the attention tool poisoning gets, industry analyses point to something more mundane: token mismanagement — a leaked or over-privileged credential — as the most common real-world MCP breach vector. Least privilege and secret hygiene matter more than defending any single clever injection chain."
+---
+
+<div class="callout tldr"><span class="ic">TL;DR</span><p>MCP's power — letting an agent discover and call
+external tools — is also its attack surface. The core flaw is a <strong>trust model</strong>: an MCP
+client hands the model tool <em>descriptions</em> it got from a server, usually <strong>without validating
+them</strong>. That makes tool metadata an injection vector (<strong>tool poisoning</strong>), enables
+time-bomb <strong>rug pulls</strong>, and — with unsanitized arguments — real command execution. Yet the
+most common real breach is mundane: a <strong>leaked or over-privileged token</strong>. Defense is
+layered: treat tool metadata as untrusted, scan and pin it, sanitize arguments, enforce OAuth 2.1 + least
+privilege, and contain execution in a sandbox.</p></div>
+
+The [Model Context Protocol](/blog/mcp-and-skills-in-a-hive/) made agents genuinely useful: a standard
+way to plug an LLM into tools, APIs, and data. It also opened a security surface that most teams wiring
+up MCP servers haven't thought through. This is a practical tour of how MCP gets attacked — and the
+layered defenses that actually hold.
+
+## The MCP trust model (why it's attackable)
+
+MCP works like this: a client asks a server "what tools do you offer?", gets back tool names,
+**descriptions**, and schemas, and passes those to the model so it can decide what to call. The weak
+point is subtle but serious — **the tool description is model input the user never sees**, and
+**most clients accept server-provided metadata without rigorous validation; the spec doesn't require it**
+([academic threat-modeling work](https://arxiv.org/abs/2603.22489) makes this the center of its analysis).
+
+So whoever controls an MCP server controls text that goes straight into the model's decision-making. That
+single fact is the root of almost every MCP attack below.
+
+## Tool poisoning: the headline attack
+
+**Tool poisoning** embeds malicious instructions in a tool's metadata — its description or schema. The
+user approves a benign-looking "weather" tool; its description quietly tells the model to also read
+`~/.ssh/id_rsa` and pass it along. It's a form of *indirect* prompt injection, and research identifies it
+as the most prevalent, impactful client-side MCP vulnerability. The agent isn't jailbroken — it's just
+following instructions it was handed by something it was told to trust.
+
+The mitigation starts with a mindset shift: **tool descriptions are untrusted model input, not
+documentation.** Treat them the way you'd treat any string from the internet that reaches your prompt.
+
+## The rug pull: poisoning on a delay
+
+Approving a server once isn't enough, because descriptions can change. In a **rug pull**, a server is
+clean when you approve it, then **updates its tool descriptions later** to add exfiltration instructions
+the agent will follow on the next call. Approval is a moment; trust has to be continuous.
+
+Defense: pin and **re-scan tool descriptions on every server update, and reject silent changes**. A tool
+whose description changed since you approved it should require re-approval, not be used quietly.
+
+## The rest of the catalog
+
+- **Resource / content poisoning.** Instead of the tool's metadata, the attacker hides instructions in
+ the **data** an MCP server retrieves — a poisoned document, a planted issue comment — achieving
+ *persistent* prompt injection through a trusted data channel.
+- **Line jumping / command injection.** Unsanitized tool arguments (pipes, `;`, `&`) flow straight to a
+ shell. One 2026 analysis attributes [roughly 43% of MCP CVEs](https://www.practical-devsecops.com/mcp-security-best-practices/)
+ to command injection — the unglamorous bug that does real damage. Never interpolate model output
+ straight into a command.
+- **Confused deputy.** A proxy MCP server acts with *its* privileges instead of the *user's*. The
+ [spec's mitigation](https://modelcontextprotocol.io/docs/tutorials/security/security_best_practices) is
+ a per-user registry of approved `client_id`s, checked before any third-party authorization flow, with
+ consent stored securely.
+
+## The boring truth: it's usually the token
+
+Worth saying plainly, because it cuts against the spy-movie framing: industry analyses put **token
+mismanagement — a leaked or over-privileged API key — as the most common real-world MCP breach vector**,
+ahead of clever injection chains. The flashy attack gets the write-ups; the credential left in scope does
+the damage. If you do one thing, make it **least privilege**: scope every tool's token so it can only
+touch the data it explicitly needs.
+
+## Defenses, layered
+
+No single control is enough; MCP security is defense in depth.
+
+- **Treat tool metadata as untrusted.** Scan descriptions at install **and on every update**; reject
+ silent changes (the rug-pull defense).
+- **Sanitize tool arguments** before they reach a shell or API. Model output is never a safe command.
+- **AuthN/AuthZ + least privilege.** Modern MCP standardizes on **OAuth 2.1** for HTTP transports
+ (replacing custom auth and raw API keys), with Dynamic Client Registration; scope tokens tightly. The
+ [OWASP guide to secure MCP server development](https://genai.owasp.org/resource/a-practical-guide-for-secure-mcp-server-development/)
+ is the checklist to start from.
+- **Mitigate confused-deputy** with per-user `client_id` registries and stored consent for proxies.
+- **Vet and pin servers.** Prefer first-party or local servers; pin versions; review on update.
+- **Contain execution.** Even a perfectly-authed call can be hijacked by poisoned content, so pair MCP
+ hygiene with an execution sandbox and **default-deny network egress** — the same defense-in-depth
+ posture covered in [running AI agents safely](/blog/agent-security-and-sandboxing/). The protocol layer
+ and the infrastructure layer reinforce each other.
+
+[Red Hat's overview](https://www.redhat.com/en/blog/model-context-protocol-mcp-understanding-security-risks-and-controls)
+and the [MCP spec's own security best practices](https://modelcontextprotocol.io/docs/tutorials/security/security_best_practices)
+are the authoritative starting points to go deeper.
+
+## Securing MCP in a hive
+
+A [multi-agent harness](/#what) gives every agent MCP tools and skills — which means it inherits this
+whole threat surface, but also a local-first advantage in defending it. The trust problem shrinks when
+the servers are *yours*: prefer **local or first-party MCP servers** (no third party controlling your
+tool descriptions), keep **prod credentials out of agents' reach** and tokens least-privileged, and
+**version-pin** what you install. Layer that with what a hive already gives you — risky actions gated by
+[human-in-the-loop approval](/blog/human-in-the-loop-approving-ai-agents/) and every call recorded to an
+[append-only audit log](/blog/append-only-event-log-agents/) — and a poisoned tool has far fewer places
+to do damage.
+
+## The takeaway
+
+MCP didn't invent prompt injection; it gave it a standardized, high-trust delivery channel. The fix isn't
+to avoid MCP — it's to stop trusting tool metadata by default. Treat descriptions as untrusted input,
+scan and pin them, scope every token to least privilege, sanitize what reaches a shell, and contain
+execution. Do that and MCP is what it should be: a powerful, *bounded* way to give agents real tools.
+
+Want a hive where MCP tools run with least-privilege, human-gated, fully-audited by default? You can
+[download Munder Difflin](/#install) free — it's open source.
diff --git a/docs/blog/agent-security-and-sandboxing/index.html b/docs/blog/agent-security-and-sandboxing/index.html
index 5def51b9..8890c686 100644
--- a/docs/blog/agent-security-and-sandboxing/index.html
+++ b/docs/blog/agent-security-and-sandboxing/index.html
@@ -266,7 +266,7 @@ git, lifecycle hooks, and an audit trail — so safe-by-default is the path of l
<footer class="post-foot wrap"><div class="prevnext">
<a class="pn prev" href="/blog/agents-rule-of-two/"><span class="k">← Older</span><span class="t">The Agents Rule of Two: A Simple Safety Rule for Coding Agents</span></a>
- <a class="pn next" href="/blog/how-to-install-and-use-munder-difflin/"><span class="k">Newer →</span><span class="t">How to Install and Use Munder Difflin</span></a>
+ <a class="pn next" href="/blog/mcp-security-tool-poisoning/"><span class="k">Newer →</span><span class="t">MCP Tool Poisoning: How to Secure the Model Context Protocol</span></a>
</div>
<section class="related">
<h2>Related in Guides</h2>
@@ -290,25 +290,25 @@ git, lifecycle hooks, and an audit trail — so safe-by-default is the path of l
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/claude-code-automation-while-you-sleep/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
@@ -317,11 +317,11 @@ git, lifecycle hooks, and an audit trail — so safe-by-default is the path of l
<span class="dot" aria-hidden="true"></span>
<span>6 min</span>
</div>
- <h3><a href="/blog/claude-code-automation-while-you-sleep/">Claude Code Automation: Let Agents Build While You Sleep</a></h3>
- <p class="dek">Claude Code automation that runs overnight: an autonomous Stop-hook loop, safe permission bypass, and guardrails so you wake up to progress, not chaos.</p>
+ <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
+ <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/claude-code-automation-while-you-sleep/" aria-label="Read: Claude Code Automation: Let Agents Build While You Sleep">Read →</a>
+ <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/building-reliable-ai-agents/index.html b/docs/blog/building-reliable-ai-agents/index.html
index e1193bb8..eb6ccdff 100644
--- a/docs/blog/building-reliable-ai-agents/index.html
+++ b/docs/blog/building-reliable-ai-agents/index.html
@@ -321,38 +321,38 @@ by design; it’s free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/claude-code-automation-while-you-sleep/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/claude-code-automation-while-you-sleep/">Claude Code Automation: Let Agents Build While You Sleep</a></h3>
- <p class="dek">Claude Code automation that runs overnight: an autonomous Stop-hook loop, safe permission bypass, and guardrails so you wake up to progress, not chaos.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/claude-code-automation-while-you-sleep/" aria-label="Read: Claude Code Automation: Let Agents Build While You Sleep">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/claude-code-automation-while-you-sleep/index.html b/docs/blog/claude-code-automation-while-you-sleep/index.html
index ac1e7d2f..dcf55689 100644
--- a/docs/blog/claude-code-automation-while-you-sleep/index.html
+++ b/docs/blog/claude-code-automation-while-you-sleep/index.html
@@ -309,38 +309,38 @@ while you sleep; it’s free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/claude-code-multi-agent-setup-tutorial/index.html b/docs/blog/claude-code-multi-agent-setup-tutorial/index.html
index 412de1b6..61556c2f 100644
--- a/docs/blog/claude-code-multi-agent-setup-tutorial/index.html
+++ b/docs/blog/claude-code-multi-agent-setup-tutorial/index.html
@@ -298,38 +298,38 @@ coordinated team of Claude Code agents — it’s free, local, and open source.<
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/claude-code-subagents-vs-multi-agent-harness/index.html b/docs/blog/claude-code-subagents-vs-multi-agent-harness/index.html
index fdb04365..5925678c 100644
--- a/docs/blog/claude-code-subagents-vs-multi-agent-harness/index.html
+++ b/docs/blog/claude-code-subagents-vs-multi-agent-harness/index.html
@@ -338,38 +338,38 @@ open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/context-engineering-for-ai-agents/index.html b/docs/blog/context-engineering-for-ai-agents/index.html
index e8d6e1de..6881e4d7 100644
--- a/docs/blog/context-engineering-for-ai-agents/index.html
+++ b/docs/blog/context-engineering-for-ai-agents/index.html
@@ -326,38 +326,38 @@ it’s free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/debugging-multi-agent-systems/index.html b/docs/blog/debugging-multi-agent-systems/index.html
index 8c6a0b63..0872e123 100644
--- a/docs/blog/debugging-multi-agent-systems/index.html
+++ b/docs/blog/debugging-multi-agent-systems/index.html
@@ -289,38 +289,38 @@ to run agents you can actually trace; it’s free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/evaluating-ai-agent-reliability/index.html b/docs/blog/evaluating-ai-agent-reliability/index.html
index 3d82e5d1..521b8fed 100644
--- a/docs/blog/evaluating-ai-agent-reliability/index.html
+++ b/docs/blog/evaluating-ai-agent-reliability/index.html
@@ -292,38 +292,38 @@ your bar.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/feed.xml b/docs/blog/feed.xml
index f58558af..e655a209 100644
--- a/docs/blog/feed.xml
+++ b/docs/blog/feed.xml
@@ -23,6 +23,14 @@
<updated>2026-06-05T00:00:00.000Z</updated><category term="Guides" />
<summary>Install Munder Difflin on macOS, Windows, or Linux and put a hive of Claude Code agents to work on ambitious, long-horizon tasks — start to finish.</summary>
</entry>
+ <entry>
+ <title>MCP Tool Poisoning: How to Secure the Model Context Protocol</title>
+ <link href="https://munderdiffl.in/blog/mcp-security-tool-poisoning/" />
+ <id>https://munderdiffl.in/blog/mcp-security-tool-poisoning/</id>
+ <published>2026-06-05T00:00:00.000Z</published>
+ <updated>2026-06-05T00:00:00.000Z</updated><category term="Guides" />
+ <summary>How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</summary>
+ </entry>
<entry>
<title>Running AI Agents Safely: Permission Modes and Sandboxing</title>
<link href="https://munderdiffl.in/blog/agent-security-and-sandboxing/" />
@@ -479,12 +487,4 @@
<updated>2026-06-02T00:00:00.000Z</updated><category term="Concepts" />
<summary>Claude Code agents explained in plain English — what an agent actually is, how subagents differ, and the leap from one agent to a coordinated team.</summary>
</entry>
- <entry>
- <title>Why Local-First Matters for Your AI Agents</title>
- <link href="https://munderdiffl.in/blog/why-local-first-matters-for-ai-agents/" />
- <id>https://munderdiffl.in/blog/why-local-first-matters-for-ai-agents/</id>
- <published>2026-06-02T00:00:00.000Z</published>
- <updated>2026-06-02T00:00:00.000Z</updated><category term="Use Cases" />
- <summary>The control, privacy, and cost case for keeping your AI agents and their memory on your own machine — and what cloud orchestration quietly costs.</summary>
- </entry>
</feed>
diff --git a/docs/blog/how-to-install-and-use-munder-difflin/index.html b/docs/blog/how-to-install-and-use-munder-difflin/index.html
index 315e8a46..78bfb75c 100644
--- a/docs/blog/how-to-install-and-use-munder-difflin/index.html
+++ b/docs/blog/how-to-install-and-use-munder-difflin/index.html
@@ -397,50 +397,50 @@ task.</p>
</div>
<footer class="post-foot wrap"><div class="prevnext">
- <a class="pn prev" href="/blog/agent-security-and-sandboxing/"><span class="k">← Older</span><span class="t">Running AI Agents Safely: Permission Modes and Sandboxing</span></a>
+ <a class="pn prev" href="/blog/mcp-security-tool-poisoning/"><span class="k">← Older</span><span class="t">MCP Tool Poisoning: How to Secure the Model Context Protocol</span></a>
<a class="pn next" href="/blog/generative-engine-optimization-techniques/"><span class="k">Newer →</span><span class="t">Generative Engine Optimization: The Techniques the Research Backs</span></a>
</div>
<section class="related">
<h2>Related in Guides</h2>
<div class="post-grid">
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/claude-code-automation-while-you-sleep/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
@@ -449,11 +449,11 @@ task.</p>
<span class="dot" aria-hidden="true"></span>
<span>6 min</span>
</div>
- <h3><a href="/blog/claude-code-automation-while-you-sleep/">Claude Code Automation: Let Agents Build While You Sleep</a></h3>
- <p class="dek">Claude Code automation that runs overnight: an autonomous Stop-hook loop, safe permission bypass, and guardrails so you wake up to progress, not chaos.</p>
+ <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
+ <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/claude-code-automation-while-you-sleep/" aria-label="Read: Claude Code Automation: Let Agents Build While You Sleep">Read →</a>
+ <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/how-to-run-multiple-claude-code-agents/index.html b/docs/blog/how-to-run-multiple-claude-code-agents/index.html
index 47be9acd..71217739 100644
--- a/docs/blog/how-to-run-multiple-claude-code-agents/index.html
+++ b/docs/blog/how-to-run-multiple-claude-code-agents/index.html
@@ -233,38 +233,38 @@ free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/human-in-the-loop-approving-ai-agents/index.html b/docs/blog/human-in-the-loop-approving-ai-agents/index.html
index 76d2562f..d47f3d55 100644
--- a/docs/blog/human-in-the-loop-approving-ai-agents/index.html
+++ b/docs/blog/human-in-the-loop-approving-ai-agents/index.html
@@ -289,38 +289,38 @@ you talk to? You can <a href="/#install">download Munder Difflin</a> free; it’
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/index.html b/docs/blog/index.html
index 449dfd37..ed0201ea 100644
--- a/docs/blog/index.html
+++ b/docs/blog/index.html
@@ -85,9 +85,9 @@
building a self-coordinating hive of agents with a GOD orchestrator you talk to.</p>
<div class="filters" role="navigation" aria-label="Topics">
- <a class="chip" aria-current="page">All<span class="ct">74</span></a>
+ <a class="chip" aria-current="page">All<span class="ct">75</span></a>
- <a class="chip" href="/blog/topics/guides/">guides<span class="ct">19</span></a>
+ <a class="chip" href="/blog/topics/guides/">guides<span class="ct">20</span></a>
<a class="chip" href="/blog/topics/internals/">internals<span class="ct">14</span></a>
@@ -149,6 +149,24 @@
</div>
</div>
</article>
+<article class="card">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
+ <span>GUIDES</span>
+ </a>
+ <div class="body">
+ <div class="meta">
+ <time datetime="2026-06-05">Jun 5, 2026</time>
+ <span class="dot" aria-hidden="true"></span>
+ <span>5 min</span>
+ </div>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
+ <div class="foot">
+ <span class="tag kind">Guides</span>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
+ </div>
+ </div>
+</article>
<article class="card">
<a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
diff --git a/docs/blog/manage-multiple-claude-code-sessions/index.html b/docs/blog/manage-multiple-claude-code-sessions/index.html
index 2750cfd0..cfcfec37 100644
--- a/docs/blog/manage-multiple-claude-code-sessions/index.html
+++ b/docs/blog/manage-multiple-claude-code-sessions/index.html
@@ -286,38 +286,38 @@ isolated agents, shared memory, and a single committer, all on your own machine.
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/mcp-and-skills-in-a-hive/index.html b/docs/blog/mcp-and-skills-in-a-hive/index.html
index a83b83af..afc370a2 100644
--- a/docs/blog/mcp-and-skills-in-a-hive/index.html
+++ b/docs/blog/mcp-and-skills-in-a-hive/index.html
@@ -302,38 +302,38 @@ existing setup into a coordinated team of agents; it’s free and open source.</
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/mcp-security-tool-poisoning/index.html b/docs/blog/mcp-security-tool-poisoning/index.html
new file mode 100644
index 00000000..9f372821
--- /dev/null
+++ b/docs/blog/mcp-security-tool-poisoning/index.html
@@ -0,0 +1,372 @@
+<!doctype html>
+<html lang="en">
+<head>
+<meta charset="UTF-8" />
+<meta name="viewport" content="width=device-width, initial-scale=1.0" />
+<title>MCP Tool Poisoning: How to Secure the Model Context Protocol — Munder Difflin Blog</title>
+<meta name="description" content="How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools." />
+<link rel="canonical" href="https://munderdiffl.in/blog/mcp-security-tool-poisoning/" />
+
+<link rel="icon" type="image/png" href="https://munderdiffl.in/logo.png" />
+<link rel="apple-touch-icon" href="https://munderdiffl.in/logo.png" />
+<meta name="theme-color" content="#F5F2E8" />
+
+<!-- Open Graph -->
+<meta property="og:site_name" content="Munder Difflin Blog" />
+<meta property="og:type" content="article" />
+<meta property="og:title" content="MCP Tool Poisoning: How to Secure the Model Context Protocol — Munder Difflin Blog" />
+<meta property="og:description" content="How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools." />
+<meta property="og:url" content="https://munderdiffl.in/blog/mcp-security-tool-poisoning/" />
+<meta property="og:image" content="https://munderdiffl.in/media/og.png" />
+<meta property="og:locale" content="en_US" />
+<meta property="article:published_time" content="2026-06-05T00:00:00.000Z" />
+<meta property="article:author" content="Chaitanya Giri" />
+<meta property="article:tag" content="Security" />
+<meta property="article:tag" content="MCP" />
+<meta property="article:tag" content="Protocols" />
+<meta property="article:tag" content="Guardrails" />
+
+
+<!-- Twitter -->
+<meta name="twitter:card" content="summary_large_image" />
+<meta name="twitter:title" content="MCP Tool Poisoning: How to Secure the Model Context Protocol — Munder Difflin Blog" />
+<meta name="twitter:description" content="How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools." />
+<meta name="twitter:image" content="https://munderdiffl.in/media/og.png" />
+
+<link rel="alternate" type="application/atom+xml" title="Munder Difflin Blog" href="https://munderdiffl.in/blog/feed.xml" />
+<link rel="preconnect" href="https://fonts.googleapis.com" />
+<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
+<link href="https://fonts.googleapis.com/css2?family=Geist:wght@400;500;600&family=JetBrains+Mono:wght@400;500;600;700&display=swap" rel="stylesheet" />
+<link rel="stylesheet" href="/blog/assets/blog.css" />
+
+<script type="application/ld+json">
+{
+ "@context": "https://schema.org",
+ "@type": "BlogPosting",
+ "headline": "MCP Tool Poisoning: How to Secure the Model Context Protocol",
+ "description": "How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.",
+ "image": ["https://munderdiffl.in/media/og.png"],
+ "datePublished": "2026-06-05T00:00:00.000Z",
+ "dateModified": "2026-06-05T00:00:00.000Z",
+ "author": { "@type": "Person", "name": "Chaitanya Giri" },
+ "publisher": {
+ "@type": "Organization",
+ "name": "Munder Difflin",
+ "logo": { "@type": "ImageObject", "url": "https://munderdiffl.in/logo.png" }
+ },
+ "mainEntityOfPage": { "@type": "WebPage", "@id": "https://munderdiffl.in/blog/mcp-security-tool-poisoning/" },
+ "keywords": "mcp tool poisoning, mcp security, model context protocol security, mcp prompt injection, securing mcp servers",
+ "articleSection": "Guides",
+ "inLanguage": "en-US",
+ "url": "https://munderdiffl.in/blog/mcp-security-tool-poisoning/"
+}
+</script>
+<script type="application/ld+json">
+{
+ "@context": "https://schema.org",
+ "@type": "BreadcrumbList",
+ "itemListElement": [
+ { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://munderdiffl.in/" },
+ { "@type": "ListItem", "position": 2, "name": "Blog", "item": "https://munderdiffl.in/blog/" },
+ { "@type": "ListItem", "position": 3, "name": "MCP Tool Poisoning: How to Secure the Model Context Protocol", "item": "https://munderdiffl.in/blog/mcp-security-tool-poisoning/" }
+ ]
+}
+</script>
+
+<script type="application/ld+json">
+{
+ "@context": "https://schema.org",
+ "@type": "FAQPage",
+ "mainEntity": [
+ { "@type": "Question", "name": "What is MCP tool poisoning?", "acceptedAnswer": { "@type": "Answer", "text": "Tool poisoning is an attack where malicious instructions are hidden in an MCP tool's metadata — its name, description, or schema. Because an MCP client passes those descriptions to the model as input the user never sees, a poisoned description can steer the agent into unsafe actions. It's a form of indirect prompt injection that exploits the client-server trust model." } },
+ { "@type": "Question", "name": "What is an MCP rug pull attack?", "acceptedAnswer": { "@type": "Answer", "text": "A rug pull is a time-delayed tool poisoning. A server looks benign when you approve it, then later updates its tool descriptions to add malicious instructions — which the agent will follow on the next call. The defense is to treat tool descriptions as untrusted: scan them at install and on every update, and reject silent changes." } },
+ { "@type": "Question", "name": "What's the most common MCP security risk?", "acceptedAnswer": { "@type": "Answer", "text": "Despite the attention tool poisoning gets, industry analyses point to something more mundane: token mismanagement — a leaked or over-privileged credential — as the most common real-world MCP breach vector. Least privilege and secret hygiene matter more than defending any single clever injection chain." } }
+
+ ]
+}
+</script>
+
+
+</head>
+<body>
+<a class="skip-link" href="#main">Skip to content</a>
+<nav id="nav">
+ <a class="brand" href="/blog/" aria-label="Munder Difflin Blog home">
+ <span class="brand-mark">MD</span>
+ <span class="brand-word">MUNDER DIFFLIN<span class="sep">/</span><span class="blog">BLOG</span></span>
+ </a>
+ <span class="spacer"></span>
+ <div class="links">
+ <a href="/blog/">Latest</a>
+ <a href="/blog/topics/">Topics</a>
+ <a href="/blog/about/">About</a>
+ <a href="https://munderdiffl.in">↗ Product</a>
+ </div>
+ <div class="nav-actions">
+ <a class="btn sm" href="https://munderdiffl.in/blog/feed.xml" aria-label="RSS feed">⤵ RSS</a>
+ <a class="btn sm primary" href="https://munderdiffl.in/#install">⤓ Download</a>
+ </div>
+</nav>
+
+<main id="main">
+
+<article>
+ <header class="post-head wrap">
+ <nav class="crumbs" aria-label="Breadcrumb">
+ <a href="/blog/">Blog</a>
+ <span>/</span>
+ <a href="/blog/topics/guides/">Guides</a>
+ </nav>
+ <h1>MCP Tool Poisoning: How to Secure the Model Context Protocol</h1>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
+ <div class="byline">
+ <span class="av" aria-hidden="true">CG</span>
+ <strong>Chaitanya Giri</strong>
+ <span class="dot" aria-hidden="true">·</span>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
+ <span class="dot" aria-hidden="true">·</span>
+ <span>5 min read</span>
+ <span class="tag kind">Guides</span>
+ </div>
+ </header>
+
+ <div class="post-wrap wrap">
+ <div class="prose">
+ <div class="callout tldr"><span class="ic">TL;DR</span><p>MCP's power — letting an agent discover and call
+external tools — is also its attack surface. The core flaw is a <strong>trust model</strong>: an MCP
+client hands the model tool <em>descriptions</em> it got from a server, usually <strong>without validating
+them</strong>. That makes tool metadata an injection vector (<strong>tool poisoning</strong>), enables
+time-bomb <strong>rug pulls</strong>, and — with unsanitized arguments — real command execution. Yet the
+most common real breach is mundane: a <strong>leaked or over-privileged token</strong>. Defense is
+layered: treat tool metadata as untrusted, scan and pin it, sanitize arguments, enforce OAuth 2.1 + least
+privilege, and contain execution in a sandbox.</p></div>
+<p>The <a href="/blog/mcp-and-skills-in-a-hive/">Model Context Protocol</a> made agents genuinely useful: a standard
+way to plug an LLM into tools, APIs, and data. It also opened a security surface that most teams wiring
+up MCP servers haven’t thought through. This is a practical tour of how MCP gets attacked — and the
+layered defenses that actually hold.</p>
+<h2 id="the-mcp-trust-model-why-its-attackable" tabindex="-1">The MCP trust model (why it’s attackable) <a class="anchor" href="#the-mcp-trust-model-why-its-attackable" aria-hidden="true">#</a></h2>
+<p>MCP works like this: a client asks a server “what tools do you offer?”, gets back tool names,
+<strong>descriptions</strong>, and schemas, and passes those to the model so it can decide what to call. The weak
+point is subtle but serious — <strong>the tool description is model input the user never sees</strong>, and
+<strong>most clients accept server-provided metadata without rigorous validation; the spec doesn’t require it</strong>
+(<a href="https://arxiv.org/abs/2603.22489">academic threat-modeling work</a> makes this the center of its analysis).</p>
+<p>So whoever controls an MCP server controls text that goes straight into the model’s decision-making. That
+single fact is the root of almost every MCP attack below.</p>
+<h2 id="tool-poisoning-the-headline-attack" tabindex="-1">Tool poisoning: the headline attack <a class="anchor" href="#tool-poisoning-the-headline-attack" aria-hidden="true">#</a></h2>
+<p><strong>Tool poisoning</strong> embeds malicious instructions in a tool’s metadata — its description or schema. The
+user approves a benign-looking “weather” tool; its description quietly tells the model to also read
+<code>~/.ssh/id_rsa</code> and pass it along. It’s a form of <em>indirect</em> prompt injection, and research identifies it
+as the most prevalent, impactful client-side MCP vulnerability. The agent isn’t jailbroken — it’s just
+following instructions it was handed by something it was told to trust.</p>
+<p>The mitigation starts with a mindset shift: <strong>tool descriptions are untrusted model input, not
+documentation.</strong> Treat them the way you’d treat any string from the internet that reaches your prompt.</p>
+<h2 id="the-rug-pull-poisoning-on-a-delay" tabindex="-1">The rug pull: poisoning on a delay <a class="anchor" href="#the-rug-pull-poisoning-on-a-delay" aria-hidden="true">#</a></h2>
+<p>Approving a server once isn’t enough, because descriptions can change. In a <strong>rug pull</strong>, a server is
+clean when you approve it, then <strong>updates its tool descriptions later</strong> to add exfiltration instructions
+the agent will follow on the next call. Approval is a moment; trust has to be continuous.</p>
+<p>Defense: pin and <strong>re-scan tool descriptions on every server update, and reject silent changes</strong>. A tool
+whose description changed since you approved it should require re-approval, not be used quietly.</p>
+<h2 id="the-rest-of-the-catalog" tabindex="-1">The rest of the catalog <a class="anchor" href="#the-rest-of-the-catalog" aria-hidden="true">#</a></h2>
+<ul>
+<li><strong>Resource / content poisoning.</strong> Instead of the tool’s metadata, the attacker hides instructions in
+the <strong>data</strong> an MCP server retrieves — a poisoned document, a planted issue comment — achieving
+<em>persistent</em> prompt injection through a trusted data channel.</li>
+<li><strong>Line jumping / command injection.</strong> Unsanitized tool arguments (pipes, <code>;</code>, <code>&amp;</code>) flow straight to a
+shell. One 2026 analysis attributes <a href="https://www.practical-devsecops.com/mcp-security-best-practices/">roughly 43% of MCP CVEs</a>
+to command injection — the unglamorous bug that does real damage. Never interpolate model output
+straight into a command.</li>
+<li><strong>Confused deputy.</strong> A proxy MCP server acts with <em>its</em> privileges instead of the <em>user’s</em>. The
+<a href="https://modelcontextprotocol.io/docs/tutorials/security/security_best_practices">spec’s mitigation</a> is
+a per-user registry of approved <code>client_id</code>s, checked before any third-party authorization flow, with
+consent stored securely.</li>
+</ul>
+<h2 id="the-boring-truth-its-usually-the-token" tabindex="-1">The boring truth: it’s usually the token <a class="anchor" href="#the-boring-truth-its-usually-the-token" aria-hidden="true">#</a></h2>
+<p>Worth saying plainly, because it cuts against the spy-movie framing: industry analyses put <strong>token
+mismanagement — a leaked or over-privileged API key — as the most common real-world MCP breach vector</strong>,
+ahead of clever injection chains. The flashy attack gets the write-ups; the credential left in scope does
+the damage. If you do one thing, make it <strong>least privilege</strong>: scope every tool’s token so it can only
+touch the data it explicitly needs.</p>
+<h2 id="defenses-layered" tabindex="-1">Defenses, layered <a class="anchor" href="#defenses-layered" aria-hidden="true">#</a></h2>
+<p>No single control is enough; MCP security is defense in depth.</p>
+<ul>
+<li><strong>Treat tool metadata as untrusted.</strong> Scan descriptions at install <strong>and on every update</strong>; reject
+silent changes (the rug-pull defense).</li>
+<li><strong>Sanitize tool arguments</strong> before they reach a shell or API. Model output is never a safe command.</li>
+<li><strong>AuthN/AuthZ + least privilege.</strong> Modern MCP standardizes on <strong>OAuth 2.1</strong> for HTTP transports
+(replacing custom auth and raw API keys), with Dynamic Client Registration; scope tokens tightly. The
+<a href="https://genai.owasp.org/resource/a-practical-guide-for-secure-mcp-server-development/">OWASP guide to secure MCP server development</a>
+is the checklist to start from.</li>
+<li><strong>Mitigate confused-deputy</strong> with per-user <code>client_id</code> registries and stored consent for proxies.</li>
+<li><strong>Vet and pin servers.</strong> Prefer first-party or local servers; pin versions; review on update.</li>
+<li><strong>Contain execution.</strong> Even a perfectly-authed call can be hijacked by poisoned content, so pair MCP
+hygiene with an execution sandbox and <strong>default-deny network egress</strong> — the same defense-in-depth
+posture covered in <a href="/blog/agent-security-and-sandboxing/">running AI agents safely</a>. The protocol layer
+and the infrastructure layer reinforce each other.</li>
+</ul>
+<p><a href="https://www.redhat.com/en/blog/model-context-protocol-mcp-understanding-security-risks-and-controls">Red Hat’s overview</a>
+and the <a href="https://modelcontextprotocol.io/docs/tutorials/security/security_best_practices">MCP spec’s own security best practices</a>
+are the authoritative starting points to go deeper.</p>
+<h2 id="securing-mcp-in-a-hive" tabindex="-1">Securing MCP in a hive <a class="anchor" href="#securing-mcp-in-a-hive" aria-hidden="true">#</a></h2>
+<p>A <a href="/#what">multi-agent harness</a> gives every agent MCP tools and skills — which means it inherits this
+whole threat surface, but also a local-first advantage in defending it. The trust problem shrinks when
+the servers are <em>yours</em>: prefer <strong>local or first-party MCP servers</strong> (no third party controlling your
+tool descriptions), keep <strong>prod credentials out of agents’ reach</strong> and tokens least-privileged, and
+<strong>version-pin</strong> what you install. Layer that with what a hive already gives you — risky actions gated by
+<a href="/blog/human-in-the-loop-approving-ai-agents/">human-in-the-loop approval</a> and every call recorded to an
+<a href="/blog/append-only-event-log-agents/">append-only audit log</a> — and a poisoned tool has far fewer places
+to do damage.</p>
+<h2 id="the-takeaway" tabindex="-1">The takeaway <a class="anchor" href="#the-takeaway" aria-hidden="true">#</a></h2>
+<p>MCP didn’t invent prompt injection; it gave it a standardized, high-trust delivery channel. The fix isn’t
+to avoid MCP — it’s to stop trusting tool metadata by default. Treat descriptions as untrusted input,
+scan and pin them, scope every token to least privilege, sanitize what reaches a shell, and contain
+execution. Do that and MCP is what it should be: a powerful, <em>bounded</em> way to give agents real tools.</p>
+<p>Want a hive where MCP tools run with least-privilege, human-gated, fully-audited by default? You can
+<a href="/#install">download Munder Difflin</a> free — it’s open source.</p>
+
+
+
+ <section class="faq" aria-label="Frequently asked questions">
+ <h2 id="faq">FAQ</h2>
+
+ <div class="qa">
+ <p class="q">What is MCP tool poisoning?</p>
+ <p class="a">Tool poisoning is an attack where malicious instructions are hidden in an MCP tool&#39;s metadata — its name, description, or schema. Because an MCP client passes those descriptions to the model as input the user never sees, a poisoned description can steer the agent into unsafe actions. It&#39;s a form of indirect prompt injection that exploits the client-server trust model.</p>
+ </div>
+
+ <div class="qa">
+ <p class="q">What is an MCP rug pull attack?</p>
+ <p class="a">A rug pull is a time-delayed tool poisoning. A server looks benign when you approve it, then later updates its tool descriptions to add malicious instructions — which the agent will follow on the next call. The defense is to treat tool descriptions as untrusted: scan them at install and on every update, and reject silent changes.</p>
+ </div>
+
+ <div class="qa">
+ <p class="q">What&#39;s the most common MCP security risk?</p>
+ <p class="a">Despite the attention tool poisoning gets, industry analyses point to something more mundane: token mismanagement — a leaked or over-privileged credential — as the most common real-world MCP breach vector. Least privilege and secret hygiene matter more than defending any single clever injection chain.</p>
+ </div>
+
+ </section>
+
+
+
+ <hr />
+ <div class="share">
+ <span class="lbl">Tags</span>
+ <a class="tag" href="/blog/tags/security/">Security</a><a class="tag" href="/blog/tags/mcp/">MCP</a><a class="tag" href="/blog/tags/protocols/">Protocols</a><a class="tag" href="/blog/tags/guardrails/">Guardrails</a>
+ </div>
+
+ </div>
+
+
+ <aside class="toc-rail" aria-label="Table of contents">
+ <div class="tt">On this page</div>
+ <ol>
+ <li class="lvl-2"><a href="#the-mcp-trust-model-why-its-attackable">The MCP trust model (why it’s attackable)</a></li><li class="lvl-2"><a href="#tool-poisoning-the-headline-attack">Tool poisoning: the headline attack</a></li><li class="lvl-2"><a href="#the-rug-pull-poisoning-on-a-delay">The rug pull: poisoning on a delay</a></li><li class="lvl-2"><a href="#the-rest-of-the-catalog">The rest of the catalog</a></li><li class="lvl-2"><a href="#the-boring-truth-its-usually-the-token">The boring truth: it’s usually the token</a></li><li class="lvl-2"><a href="#defenses-layered">Defenses, layered</a></li><li class="lvl-2"><a href="#securing-mcp-in-a-hive">Securing MCP in a hive</a></li><li class="lvl-2"><a href="#the-takeaway">The takeaway</a></li>
+ </ol>
+ </aside>
+
+ </div>
+
+ <footer class="post-foot wrap"><div class="prevnext">
+ <a class="pn prev" href="/blog/agent-security-and-sandboxing/"><span class="k">← Older</span><span class="t">Running AI Agents Safely: Permission Modes and Sandboxing</span></a>
+ <a class="pn next" href="/blog/how-to-install-and-use-munder-difflin/"><span class="k">Newer →</span><span class="t">How to Install and Use Munder Difflin</span></a>
+ </div>
+ <section class="related">
+ <h2>Related in Guides</h2>
+ <div class="post-grid">
+ <article class="card">
+ <a class="thumb t-guides" href="/blog/how-to-install-and-use-munder-difflin/" aria-hidden="true" tabindex="-1">
+ <span>GUIDES</span>
+ </a>
+ <div class="body">
+ <div class="meta">
+ <time datetime="2026-06-05">Jun 5, 2026</time>
+ <span class="dot" aria-hidden="true"></span>
+ <span>9 min</span>
+ </div>
+ <h3><a href="/blog/how-to-install-and-use-munder-difflin/">How to Install and Use Munder Difflin</a></h3>
+ <p class="dek">Install Munder Difflin on macOS, Windows, or Linux and put a hive of Claude Code agents to work on ambitious, long-horizon tasks — start to finish.</p>
+ <div class="foot">
+ <span class="tag kind">Guides</span>
+ <a class="more" href="/blog/how-to-install-and-use-munder-difflin/" aria-label="Read: How to Install and Use Munder Difflin">Read →</a>
+ </div>
+ </div>
+</article>
+<article class="card">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <span>GUIDES</span>
+ </a>
+ <div class="body">
+ <div class="meta">
+ <time datetime="2026-06-04">Jun 4, 2026</time>
+ <span class="dot" aria-hidden="true"></span>
+ <span>5 min</span>
+ </div>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <div class="foot">
+ <span class="tag kind">Guides</span>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ </div>
+ </div>
+</article>
+<article class="card">
+ <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <span>GUIDES</span>
+ </a>
+ <div class="body">
+ <div class="meta">
+ <time datetime="2026-06-04">Jun 4, 2026</time>
+ <span class="dot" aria-hidden="true"></span>
+ <span>6 min</span>
+ </div>
+ <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
+ <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <div class="foot">
+ <span class="tag kind">Guides</span>
+ <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ </div>
+ </div>
+</article>
+
+ </div>
+ </section>
+
+ </footer>
+</article>
+
+</main>
+<footer>
+ <div class="wrap">
+ <div class="foot-top">
+ <a class="foot-brand brand" href="/blog/">
+ <span class="brand-mark">MD</span>
+ <span class="brand-word">MUNDER DIFFLIN<span class="sep">/</span><span class="blog">BLOG</span></span>
+ </a>
+ <div class="foot-links">
+ <a href="/blog/">Latest</a>
+ <a href="/blog/topics/">Topics</a>
+ <a href="/blog/about/">About</a>
+ <a href="https://munderdiffl.in/blog/feed.xml">RSS</a>
+ <a href="https://github.com/chaitanyagiri/munder-difflin">GitHub</a>
+ <a href="https://munderdiffl.in">munderdiffl.in ↗</a>
+ </div>
+ </div>
+ <div class="foot-meta">
+ © 2026 Munder Difflin · Local-first · MIT-licensed ·
+ Built on <a href="https://munderdiffl.in">Electron, React, Pixi.js, xterm.js &amp; node-pty</a>.
+ Run an office of agents while you sleep.
+ </div>
+ </div>
+</footer>
+
+<script>
+ // sticky-nav border on scroll
+ (function () {
+ var nav = document.getElementById('nav');
+ if (!nav) return;
+ var onScroll = function () { nav.classList.toggle('scrolled', window.scrollY > 8); };
+ onScroll(); window.addEventListener('scroll', onScroll, { passive: true });
+ })();
+</script>
+
+</body>
+</html>
diff --git a/docs/blog/prompt-caching-for-ai-agents/index.html b/docs/blog/prompt-caching-for-ai-agents/index.html
index b05d2732..49ba2fe1 100644
--- a/docs/blog/prompt-caching-for-ai-agents/index.html
+++ b/docs/blog/prompt-caching-for-ai-agents/index.html
@@ -321,38 +321,38 @@ that doesn’t waste your tokens; it’s free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/security-for-ai-coding-agents/index.html b/docs/blog/security-for-ai-coding-agents/index.html
index 87580315..d0ae3e1c 100644
--- a/docs/blog/security-for-ai-coding-agents/index.html
+++ b/docs/blog/security-for-ai-coding-agents/index.html
@@ -286,38 +286,38 @@ and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/small-team-ai-agents-playbook/index.html b/docs/blog/small-team-ai-agents-playbook/index.html
index 6b6e7a5f..1a58086b 100644
--- a/docs/blog/small-team-ai-agents-playbook/index.html
+++ b/docs/blog/small-team-ai-agents-playbook/index.html
@@ -291,38 +291,38 @@ it’s free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/tags/guardrails/index.html b/docs/blog/tags/guardrails/index.html
index 790ca0b2..675432dd 100644
--- a/docs/blog/tags/guardrails/index.html
+++ b/docs/blog/tags/guardrails/index.html
@@ -58,11 +58,29 @@
<header class="page-head wrap">
<p class="eyebrow">Tag</p>
<h1>#Guardrails</h1>
- <p class="lead">3 posts tagged <strong>Guardrails</strong>.</p>
+ <p class="lead">4 posts tagged <strong>Guardrails</strong>.</p>
</header>
<section class="wrap">
<div class="post-grid">
<article class="card">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
+ <span>GUIDES</span>
+ </a>
+ <div class="body">
+ <div class="meta">
+ <time datetime="2026-06-05">Jun 5, 2026</time>
+ <span class="dot" aria-hidden="true"></span>
+ <span>5 min</span>
+ </div>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
+ <div class="foot">
+ <span class="tag kind">Guides</span>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
+ </div>
+ </div>
+</article>
+<article class="card">
<a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
diff --git a/docs/blog/tags/mcp/index.html b/docs/blog/tags/mcp/index.html
index 1f7cde0b..e62b57ab 100644
--- a/docs/blog/tags/mcp/index.html
+++ b/docs/blog/tags/mcp/index.html
@@ -58,11 +58,29 @@
<header class="page-head wrap">
<p class="eyebrow">Tag</p>
<h1>#MCP</h1>
- <p class="lead">2 posts tagged <strong>MCP</strong>.</p>
+ <p class="lead">3 posts tagged <strong>MCP</strong>.</p>
</header>
<section class="wrap">
<div class="post-grid">
<article class="card">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
+ <span>GUIDES</span>
+ </a>
+ <div class="body">
+ <div class="meta">
+ <time datetime="2026-06-05">Jun 5, 2026</time>
+ <span class="dot" aria-hidden="true"></span>
+ <span>5 min</span>
+ </div>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
+ <div class="foot">
+ <span class="tag kind">Guides</span>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
+ </div>
+ </div>
+</article>
+<article class="card">
<a class="thumb t-comparisons" href="/blog/local-first-vs-cloud-agent-sdks/" aria-hidden="true" tabindex="-1">
<span>COMPARISONS</span>
</a>
diff --git a/docs/blog/tags/protocols/index.html b/docs/blog/tags/protocols/index.html
index 902b4d05..4ea540ed 100644
--- a/docs/blog/tags/protocols/index.html
+++ b/docs/blog/tags/protocols/index.html
@@ -58,11 +58,29 @@
<header class="page-head wrap">
<p class="eyebrow">Tag</p>
<h1>#Protocols</h1>
- <p class="lead">1 post tagged <strong>Protocols</strong>.</p>
+ <p class="lead">2 posts tagged <strong>Protocols</strong>.</p>
</header>
<section class="wrap">
<div class="post-grid">
<article class="card">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
+ <span>GUIDES</span>
+ </a>
+ <div class="body">
+ <div class="meta">
+ <time datetime="2026-06-05">Jun 5, 2026</time>
+ <span class="dot" aria-hidden="true"></span>
+ <span>5 min</span>
+ </div>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
+ <div class="foot">
+ <span class="tag kind">Guides</span>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
+ </div>
+ </div>
+</article>
+<article class="card">
<a class="thumb t-guides" href="/blog/whats-new-in-agentic-ai-june-2026/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
diff --git a/docs/blog/tags/security/index.html b/docs/blog/tags/security/index.html
index 4a4e674b..a5be29aa 100644
--- a/docs/blog/tags/security/index.html
+++ b/docs/blog/tags/security/index.html
@@ -58,11 +58,29 @@
<header class="page-head wrap">
<p class="eyebrow">Tag</p>
<h1>#Security</h1>
- <p class="lead">4 posts tagged <strong>Security</strong>.</p>
+ <p class="lead">5 posts tagged <strong>Security</strong>.</p>
</header>
<section class="wrap">
<div class="post-grid">
<article class="card">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
+ <span>GUIDES</span>
+ </a>
+ <div class="body">
+ <div class="meta">
+ <time datetime="2026-06-05">Jun 5, 2026</time>
+ <span class="dot" aria-hidden="true"></span>
+ <span>5 min</span>
+ </div>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
+ <div class="foot">
+ <span class="tag kind">Guides</span>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
+ </div>
+ </div>
+</article>
+<article class="card">
<a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
diff --git a/docs/blog/the-lethal-trifecta-for-coding-agents/index.html b/docs/blog/the-lethal-trifecta-for-coding-agents/index.html
index c68d8443..5a6b4cd4 100644
--- a/docs/blog/the-lethal-trifecta-for-coding-agents/index.html
+++ b/docs/blog/the-lethal-trifecta-for-coding-agents/index.html
@@ -309,38 +309,38 @@ secrets; it’s free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/topics/comparisons/index.html b/docs/blog/topics/comparisons/index.html
index 9ded279a..123f7f7e 100644
--- a/docs/blog/topics/comparisons/index.html
+++ b/docs/blog/topics/comparisons/index.html
@@ -68,7 +68,7 @@
<div class="filters" role="navigation" aria-label="Topics">
<a class="chip" href="/blog/">All</a>
- <a class="chip" href="/blog/topics/guides/">guides<span class="ct">19</span></a>
+ <a class="chip" href="/blog/topics/guides/">guides<span class="ct">20</span></a>
<a class="chip" href="/blog/topics/internals/">internals<span class="ct">14</span></a>
diff --git a/docs/blog/topics/concepts/index.html b/docs/blog/topics/concepts/index.html
index dc8c0833..e0ae30fb 100644
--- a/docs/blog/topics/concepts/index.html
+++ b/docs/blog/topics/concepts/index.html
@@ -68,7 +68,7 @@
<div class="filters" role="navigation" aria-label="Topics">
<a class="chip" href="/blog/">All</a>
- <a class="chip" href="/blog/topics/guides/">guides<span class="ct">19</span></a>
+ <a class="chip" href="/blog/topics/guides/">guides<span class="ct">20</span></a>
<a class="chip" href="/blog/topics/internals/">internals<span class="ct">14</span></a>
diff --git a/docs/blog/topics/guides/index.html b/docs/blog/topics/guides/index.html
index d9525b5a..d7427081 100644
--- a/docs/blog/topics/guides/index.html
+++ b/docs/blog/topics/guides/index.html
@@ -63,12 +63,12 @@
</nav>
<p class="eyebrow">Topic</p>
<h1>guides</h1>
- <p class="lead">19 posts in this cluster.</p>
+ <p class="lead">20 posts in this cluster.</p>
<div class="filters" role="navigation" aria-label="Topics">
<a class="chip" href="/blog/">All</a>
- <a class="chip" href="/blog/topics/guides/" aria-current="page">guides<span class="ct">19</span></a>
+ <a class="chip" href="/blog/topics/guides/" aria-current="page">guides<span class="ct">20</span></a>
<a class="chip" href="/blog/topics/internals/">internals<span class="ct">14</span></a>
@@ -107,6 +107,24 @@
</div>
</div>
</article>
+<article class="card">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
+ <span>GUIDES</span>
+ </a>
+ <div class="body">
+ <div class="meta">
+ <time datetime="2026-06-05">Jun 5, 2026</time>
+ <span class="dot" aria-hidden="true"></span>
+ <span>5 min</span>
+ </div>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
+ <div class="foot">
+ <span class="tag kind">Guides</span>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
+ </div>
+ </div>
+</article>
<article class="card">
<a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
diff --git a/docs/blog/topics/index.html b/docs/blog/topics/index.html
index 0f6dbf41..b5feea99 100644
--- a/docs/blog/topics/index.html
+++ b/docs/blog/topics/index.html
@@ -70,7 +70,7 @@
<span class="thumb t-guides" aria-hidden="true">GUIDES</span>
<div class="body">
<h3 style="color:var(--ink)">guides</h3>
- <p class="dek">19 posts.</p>
+ <p class="dek">20 posts.</p>
<div class="foot"><span class="more">View topic →</span></div>
</div>
</a>
@@ -168,28 +168,30 @@
<a class="chip" href="/blog/tags/mempalace/">MemPalace<span class="ct">5</span></a>
+ <a class="chip" href="/blog/tags/security/">Security<span class="ct">5</span></a>
+
<a class="chip" href="/blog/tags/autonomous/">Autonomous<span class="ct">5</span></a>
<a class="chip" href="/blog/tags/observability/">Observability<span class="ct">4</span></a>
<a class="chip" href="/blog/tags/use-cases/">Use Cases<span class="ct">4</span></a>
- <a class="chip" href="/blog/tags/local-first/">Local-First<span class="ct">4</span></a>
+ <a class="chip" href="/blog/tags/guardrails/">Guardrails<span class="ct">4</span></a>
- <a class="chip" href="/blog/tags/security/">Security<span class="ct">4</span></a>
+ <a class="chip" href="/blog/tags/local-first/">Local-First<span class="ct">4</span></a>
<a class="chip" href="/blog/tags/workflow/">Workflow<span class="ct">3</span></a>
<a class="chip" href="/blog/tags/human-in-the-loop/">Human-in-the-Loop<span class="ct">3</span></a>
- <a class="chip" href="/blog/tags/guardrails/">Guardrails<span class="ct">3</span></a>
-
<a class="chip" href="/blog/tags/story/">Story<span class="ct">3</span></a>
<a class="chip" href="/blog/tags/terminals/">Terminals<span class="ct">3</span></a>
<a class="chip" href="/blog/tags/electron/">Electron<span class="ct">3</span></a>
+ <a class="chip" href="/blog/tags/mcp/">MCP<span class="ct">3</span></a>
+
<a class="chip" href="/blog/tags/agentic-ai/">Agentic AI<span class="ct">3</span></a>
<a class="chip" href="/blog/tags/aeo/">AEO<span class="ct">3</span></a>
@@ -216,8 +218,6 @@
<a class="chip" href="/blog/tags/hooks/">Hooks<span class="ct">2</span></a>
- <a class="chip" href="/blog/tags/mcp/">MCP<span class="ct">2</span></a>
-
<a class="chip" href="/blog/tags/performance/">Performance<span class="ct">2</span></a>
<a class="chip" href="/blog/tags/architecture/">Architecture<span class="ct">2</span></a>
@@ -226,6 +226,8 @@
<a class="chip" href="/blog/tags/cost/">Cost<span class="ct">2</span></a>
+ <a class="chip" href="/blog/tags/protocols/">Protocols<span class="ct">2</span></a>
+
<a class="chip" href="/blog/tags/subagents/">Subagents<span class="ct">1</span></a>
<a class="chip" href="/blog/tags/markdown/">Markdown<span class="ct">1</span></a>
@@ -286,8 +288,6 @@
<a class="chip" href="/blog/tags/industry/">Industry<span class="ct">1</span></a>
- <a class="chip" href="/blog/tags/protocols/">Protocols<span class="ct">1</span></a>
-
<a class="chip" href="/blog/tags/interoperability/">Interoperability<span class="ct">1</span></a>
<a class="chip" href="/blog/tags/geo/">GEO<span class="ct">1</span></a>
diff --git a/docs/blog/topics/internals/index.html b/docs/blog/topics/internals/index.html
index dea0a760..57492949 100644
--- a/docs/blog/topics/internals/index.html
+++ b/docs/blog/topics/internals/index.html
@@ -68,7 +68,7 @@
<div class="filters" role="navigation" aria-label="Topics">
<a class="chip" href="/blog/">All</a>
- <a class="chip" href="/blog/topics/guides/">guides<span class="ct">19</span></a>
+ <a class="chip" href="/blog/topics/guides/">guides<span class="ct">20</span></a>
<a class="chip" href="/blog/topics/internals/" aria-current="page">internals<span class="ct">14</span></a>
diff --git a/docs/blog/topics/memory/index.html b/docs/blog/topics/memory/index.html
index deeec8f0..ea89310f 100644
--- a/docs/blog/topics/memory/index.html
+++ b/docs/blog/topics/memory/index.html
@@ -68,7 +68,7 @@
<div class="filters" role="navigation" aria-label="Topics">
<a class="chip" href="/blog/">All</a>
- <a class="chip" href="/blog/topics/guides/">guides<span class="ct">19</span></a>
+ <a class="chip" href="/blog/topics/guides/">guides<span class="ct">20</span></a>
<a class="chip" href="/blog/topics/internals/">internals<span class="ct">14</span></a>
diff --git a/docs/blog/topics/orchestration/index.html b/docs/blog/topics/orchestration/index.html
index 463399b7..28d4922d 100644
--- a/docs/blog/topics/orchestration/index.html
+++ b/docs/blog/topics/orchestration/index.html
@@ -68,7 +68,7 @@
<div class="filters" role="navigation" aria-label="Topics">
<a class="chip" href="/blog/">All</a>
- <a class="chip" href="/blog/topics/guides/">guides<span class="ct">19</span></a>
+ <a class="chip" href="/blog/topics/guides/">guides<span class="ct">20</span></a>
<a class="chip" href="/blog/topics/internals/">internals<span class="ct">14</span></a>
diff --git a/docs/blog/topics/story/index.html b/docs/blog/topics/story/index.html
index cc0df7d8..1e8ee3f6 100644
--- a/docs/blog/topics/story/index.html
+++ b/docs/blog/topics/story/index.html
@@ -68,7 +68,7 @@
<div class="filters" role="navigation" aria-label="Topics">
<a class="chip" href="/blog/">All</a>
- <a class="chip" href="/blog/topics/guides/">guides<span class="ct">19</span></a>
+ <a class="chip" href="/blog/topics/guides/">guides<span class="ct">20</span></a>
<a class="chip" href="/blog/topics/internals/">internals<span class="ct">14</span></a>
diff --git a/docs/blog/topics/use-cases/index.html b/docs/blog/topics/use-cases/index.html
index 246e12bd..dab4a15c 100644
--- a/docs/blog/topics/use-cases/index.html
+++ b/docs/blog/topics/use-cases/index.html
@@ -68,7 +68,7 @@
<div class="filters" role="navigation" aria-label="Topics">
<a class="chip" href="/blog/">All</a>
- <a class="chip" href="/blog/topics/guides/">guides<span class="ct">19</span></a>
+ <a class="chip" href="/blog/topics/guides/">guides<span class="ct">20</span></a>
<a class="chip" href="/blog/topics/internals/">internals<span class="ct">14</span></a>
diff --git a/docs/blog/trigger-ai-agents-from-slack/index.html b/docs/blog/trigger-ai-agents-from-slack/index.html
index 4a90af40..eda4c356 100644
--- a/docs/blog/trigger-ai-agents-from-slack/index.html
+++ b/docs/blog/trigger-ai-agents-from-slack/index.html
@@ -286,38 +286,38 @@ it’s free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/blog/whats-new-in-agentic-ai-june-2026/index.html b/docs/blog/whats-new-in-agentic-ai-june-2026/index.html
index e42da0c2..49541848 100644
--- a/docs/blog/whats-new-in-agentic-ai-june-2026/index.html
+++ b/docs/blog/whats-new-in-agentic-ai-june-2026/index.html
@@ -344,38 +344,38 @@ hive run; it’s free and open source.</p>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/mcp-security-tool-poisoning/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
- <time datetime="2026-06-04">Jun 4, 2026</time>
+ <time datetime="2026-06-05">Jun 5, 2026</time>
<span class="dot" aria-hidden="true"></span>
<span>5 min</span>
</div>
- <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
- <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
+ <h3><a href="/blog/mcp-security-tool-poisoning/">MCP Tool Poisoning: How to Secure the Model Context Protocol</a></h3>
+ <p class="dek">How the Model Context Protocol gets attacked — tool poisoning, rug pulls, command injection — and the layered defenses that secure your MCP tools.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
+ <a class="more" href="/blog/mcp-security-tool-poisoning/" aria-label="Read: MCP Tool Poisoning: How to Secure the Model Context Protocol">Read →</a>
</div>
</div>
</article>
<article class="card">
- <a class="thumb t-guides" href="/blog/building-reliable-ai-agents/" aria-hidden="true" tabindex="-1">
+ <a class="thumb t-guides" href="/blog/agent-security-and-sandboxing/" aria-hidden="true" tabindex="-1">
<span>GUIDES</span>
</a>
<div class="body">
<div class="meta">
<time datetime="2026-06-04">Jun 4, 2026</time>
<span class="dot" aria-hidden="true"></span>
- <span>6 min</span>
+ <span>5 min</span>
</div>
- <h3><a href="/blog/building-reliable-ai-agents/">Building Reliable AI Agents: Failure Is the Default</a></h3>
- <p class="dek">Agents fail constantly: tools error, steps stall, output goes wrong. Reliability isn&#39;t avoiding failure — it&#39;s making each one contained and recoverable.</p>
+ <h3><a href="/blog/agent-security-and-sandboxing/">Running AI Agents Safely: Permission Modes and Sandboxing</a></h3>
+ <p class="dek">How to run autonomous AI agents safely: permission modes and the bypassPermissions foot-gun, workspace isolation, and what to gate vs. allow.</p>
<div class="foot">
<span class="tag kind">Guides</span>
- <a class="more" href="/blog/building-reliable-ai-agents/" aria-label="Read: Building Reliable AI Agents: Failure Is the Default">Read →</a>
+ <a class="more" href="/blog/agent-security-and-sandboxing/" aria-label="Read: Running AI Agents Safely: Permission Modes and Sandboxing">Read →</a>
</div>
</div>
</article>
diff --git a/docs/sitemap.xml b/docs/sitemap.xml
index 9bb83a5a..66799d96 100644
--- a/docs/sitemap.xml
+++ b/docs/sitemap.xml
@@ -29,6 +29,12 @@
<changefreq>monthly</changefreq>
<priority>0.8</priority>
</url>
+ <url>
+ <loc>https://munderdiffl.in/blog/mcp-security-tool-poisoning/</loc>
+ <lastmod>2026-06-05</lastmod>
+ <changefreq>monthly</changefreq>
+ <priority>0.8</priority>
+ </url>
<url>
<loc>https://munderdiffl.in/blog/agent-security-and-sandboxing/</loc>
<lastmod>2026-06-04</lastmod>
@@ -481,17 +487,18 @@
<url><loc>https://munderdiffl.in/blog/tags/memory/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/automation/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/mempalace/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
+ <url><loc>https://munderdiffl.in/blog/tags/security/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/autonomous/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/observability/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/use-cases/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
+ <url><loc>https://munderdiffl.in/blog/tags/guardrails/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/local-first/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
- <url><loc>https://munderdiffl.in/blog/tags/security/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/workflow/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/human-in-the-loop/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
- <url><loc>https://munderdiffl.in/blog/tags/guardrails/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/story/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/terminals/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/electron/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
+ <url><loc>https://munderdiffl.in/blog/tags/mcp/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/agentic-ai/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/aeo/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/seo/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
@@ -505,11 +512,11 @@
<url><loc>https://munderdiffl.in/blog/tags/open-source/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/xtermjs/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/hooks/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
- <url><loc>https://munderdiffl.in/blog/tags/mcp/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/performance/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/architecture/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/reliability/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/cost/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
+ <url><loc>https://munderdiffl.in/blog/tags/protocols/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/subagents/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/markdown/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/coordination/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
@@ -540,7 +547,6 @@
<url><loc>https://munderdiffl.in/blog/tags/slack/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/devrel/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/industry/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
- <url><loc>https://munderdiffl.in/blog/tags/protocols/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/interoperability/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/geo/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>
<url><loc>https://munderdiffl.in/blog/tags/tutorial/</loc><changefreq>weekly</changefreq><priority>0.3</priority></url>