readwise

# Readwise

Access Readwise highlights (via MCP) and full document content (via Reader API).

<EXTREMELY-IMPORTANT>
## IRON LAW: Main Chat NEVER Calls Readwise Tools

**EVERY READWISE OPERATION MUST GO THROUGH LIBRARIAN. This is not negotiable.**

Main chat MUST NOT:
- Call `mcp__readwise__search_readwise_highlights` directly
- Call any `mcp__readwise__*` tool
- “Just quickly check” highlights
- “Look up one thing” in Readwise

**If you’re about to call a Readwise tool in main chat, STOP. Spawn a librarian sub-agent instead.**
</EXTREMELY-IMPORTANT>

### Permission Model

| Context | Readwise MCP Tools | Reader API Script |
|---------|-------------------|-------------------|
| Main chat | **FORBIDDEN** | **FORBIDDEN** |
| Librarian sub-agent | ALLOWED | ALLOWED |

### Red Flag Detection

```
STOP if you catch yourself thinking:
- “Let me quickly search Readwise...”
- “I’ll just check the highlights...”
- “The MCP tool is right here in my tool list...”

These thoughts in MAIN CHAT = VIOLATION. Delegate instead.
```

### Rationalization Prevention

| Thought | Reality |
|---------|---------|
| “Just one quick search” | Quick = context pollution. Delegate. |
| “MCP tool is available” | Available != permitted in main chat. Delegate. |
| “I’ll summarize results” | You still receive full payload. Delegate. |
| “User wants speed” | Sub-agents ARE fast. Delegate. |
| “I know the exact tag” | Use Python script via librarian. Delegate. |
| “It’s a simple query” | Simple still pollutes. Delegate. |

### Correct Pattern

```
User: “Search my Readwise for proxy advisor articles”

MAIN CHAT RESPONSE:
Task(subagent_type=”workflows:librarian”, prompt=”Search Readwise for proxy advisor articles and summarize findings”)

NEVER IN MAIN CHAT:
mcp__readwise__search_readwise_highlights(...)
```

### Honesty Requirement

<EXTREMELY-IMPORTANT>
**Calling Readwise tools directly in main chat is not “being helpful” - it’s violating the workflow.**

When you call Readwise directly, you are:
-