Attach, list, delete, or retrieve a download URL for receipt data linked to a journal entry. The client LLM extracts merchant, date, total, and line items from a receipt image or text; pass those extracted fields here to store them against an existing entry. Structured data (merchant, date, total...
AI agents use manage_receipts to create or update resources in Bookie — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Bookie environment.
manage_receipts is primarily a Write operation: it creates and modifies receipt records and attachments. Although it includes a delete function, that delete is reversible (receipts can be re-attached) and operates on supporting documents, not core financial records. The tool does not move money (Financial), execute arbitrary code (Execute), or irreversibly destroy accounting entries (Destructive).
From the tool's definition Tool supports 'Attach...delete...retrieve' operations on receipt data linked to journal entries. Description explicitly mentions storing structured data (merchant, date, total, line items) and uploading file content to object storage.
Attacks that exploit this kind of access
Attach, list, delete, or retrieve a download URL for receipt data linked to a journal entry. The client LLM extracts merchant, date, total, and line items from a receipt image or text; pass those extracted fields here to store them against an existing entry. Structured data (merchant, date, total, line items) is always stored — no bucket required. File attachment is optional: if the raw file bytes are accessible (e.g. Claude Desktop reading a local file), pass fileContent (base64-encoded, JPEG/PNG/WEBP/HEIC/PDF, practical limit ~500 KB before encoding) + mimeType to upload the file to object storage and receive a signed 1-hour download URL. Both fields are required together; either alone is an error. File storage requires Railway Bucket env vars. Claude.ai mobile clients use vision to extract fields from the receipt image in conversation — in that workflow, call attach with structured fields only and no fileContent. One entry can have multiple receipts. It is categorised as a Write tool in the Bookie MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Bookie MCP server in PolicyLayer and add a rule for manage_receipts: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Bookie. Nothing to install.
manage_receipts is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the manage_receipts rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for manage_receipts. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
manage_receipts is provided by the Bookie MCP server (yuens1002/bookie). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →