Opencode Bedrock Doc Blocks logo

Opencode Bedrock Doc Blocks

Community
OnlyTerp
opencode-bedrock-doc-blocks

OpenCode places cachePoint on Bedrock messages containing DocumentBlocks, which produces a "nothing available to cache" error.

Overview

PublisherOnlyTerp
Repositoryprompt-cache-skills
Skill nameopencode-bedrock-doc-blocks
Stars
116
Forks
10
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by OnlyTerp on GitHub. Read the source before you install it.

Installation

Install the Opencode Bedrock Doc Blocks AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/OnlyTerp/prompt-cache-skills/tree/main/skills/opencode-bedrock-doc-blocks
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/OnlyTerp/prompt-cache-skills.git /tmp/prompt-cache-skills
mkdir -p .claude/skills
cp -r /tmp/prompt-cache-skills/skills/opencode-bedrock-doc-blocks .claude/skills/opencode-bedrock-doc-blocks
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Opencode Bedrock Doc Blocks in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Opencode Bedrock Doc Blocks on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Opencode Bedrock Doc Blocks is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

OpenCode: skip cachePoint on Bedrock DocumentBlock messages (#17300)

Target

packages/opencode/src/provider/transform.ts → applyCaching() in sst/opencode.

Open issue: #17300 — "Bedrock prompt caching error: nothing available to cache".

Symptom

applyCaching() places cachePoint markers on the last 2 non-system messages. If either of those messages contains a DocumentBlock (PDF/file attachment that Bedrock processes server-side), the cachePoint lands on content that Bedrock considers non-cacheable, producing an API error:

ValidationException: nothing available to cache

The error fails the entire request, not just the caching attempt.

Fix

Add a DocumentBlock detector and skip cachePoint placement on affected messages, falling through to the next-most-recent stable non-document message:

diff
--- a/packages/opencode/src/provider/transform.ts
+++ b/packages/opencode/src/provider/transform.ts
@@ function applyCaching(msgs, model)
+  function hasDocumentBlock(msg: ModelMessage): boolean {
+    if (!Array.isArray(msg.content)) return false;
+    return msg.content.some((b: any) =>
+      b?.type === "document" ||
+      b?.type === "file" ||
+      b?.documentBlock != null
+    );
+  }
+
+  function findLastStableNonDoc(msgs: ModelMessage[], skip: number): number {
+    for (let i = msgs.length - 1 - skip; i >= 0; i--) {
+      if (msgs[i].role !== "system" && !hasDocumentBlock(msgs[i])) {
+        return i;
+      }
+    }
+    return -1;
+  }
@@
-  // Apply to last 2 non-system messages
-  const nonSystemMsgs = msgs.filter(m => m.role !== "system").slice(-2)
-  nonSystemMsgs.forEach(msg => {
-    msg.providerOptions = { /* cachePoint etc. */ }
-  })
+  // For Bedrock, skip DocumentBlock-containing messages — they
+  // can't be cached and produce a hard error.
+  const isBedrock = model.providerID.includes("bedrock");
+  if (isBedrock) {
+    const lastStable = findLastStableNonDoc(msgs, 0);
+    const prevStable = findLastStableNonDoc(msgs, msgs.length - 1 - lastStable + 1);
+    [lastStable, prevStable].forEach(idx => {
+      if (idx < 0) return;
+      msgs[idx].providerOptions = { bedrock: { cachePoint: { type: "default" } } };
+    });
+  } else {
+    // Existing behavior for non-Bedrock providers
+    const nonSystemMsgs = msgs.filter(m => m.role !== "system").slice(-2);
+    nonSystemMsgs.forEach(msg => {
+      msg.providerOptions = { /* cachePoint etc. */ };
+    });
+  }

Verify

  1. Configure OpenCode with a Bedrock Claude model.
  2. Attach a PDF or text document to a chat message.
  3. Send a follow-up prompt referencing the document.
  4. Before fix: Bedrock returns 400 ValidationException.
  5. After fix: request succeeds; cachePoint lands on the most recent non-document message; response shows cacheReadInputTokenCount > 0 on the second identical follow-up.

Background

Bedrock's cachePoint requires that the prefix it caches contains cacheable content. DocumentBlocks are processed server-side and don't count toward the cacheable token budget — placing a cachePoint on a message where the only content is a document produces the "nothing to cache" error.

The fix is to detect document-containing messages and route the breakpoint to the next-most-stable text-bearing message.

See docs/concepts/bedrock.md. Full audit: audits/opencode.md.

Frequently asked questions

What does the Opencode Bedrock Doc Blocks AI skill do?

OpenCode places cachePoint on Bedrock messages containing DocumentBlocks, which produces a "nothing available to cache" error.

Why use Opencode Bedrock Doc Blocks on TypingMind?

Because you install it once and use it with any model. Opencode Bedrock Doc Blocks is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Opencode Bedrock Doc Blocks in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/OnlyTerp/prompt-cache-skills/tree/main/skills/opencode-bedrock-doc-blocks. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Opencode Bedrock Doc Blocks?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Opencode Bedrock Doc Blocks?

As many as you like. As long as a model supports skills, you can use Opencode Bedrock Doc Blocks with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Opencode Bedrock Doc Blocks AI skill free?

It is published on GitHub by OnlyTerp. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇