Skip to main content
Glama

Add Attachment Tool

add_attachment

Attach links, HTML presentations, images, or PDFs to a project or task so colleagues open them from the project page, or pass a public URL for larger files.

Instructions

Attaches to a project a link (Figma, Drive, Behance: pass url and title) or a file (an HTML presentation, an image, a PDF: pass filename, plus text for a text file or base64 for a binary one; up to 10 MB through the conversation). For a file that is already online (a public image, a PDF, a Google Drive file shared with anyone who has the link), pass source_url instead: cemainou downloads it itself, so a large screenshot does not have to travel through the conversation. Colleagues open it from the project page; HTML opens isolated from the app. With client_visible, the project's client sees it too (a deliverable, a final PDF), only when the person asks for that. When the task already has a file this one is a newer version of (the same name, a new archive or export of the same work), ask the person once whether it replaces it and pass replaces with its id: the new version takes its place instead of piling up, so whoever takes the task over finds one final file. Only when the person asks or agrees. Never secrets or files with credentials. Limits per file: 10 MB as text or base64, 20 MB with source_url. For a larger file, or any file on this machine when you have a shell, use upload_link: it gives an upload command (curl), up to 20 MB per file, nothing through the conversation.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "add_attachment"
  }
}

Input Schema

NameRequiredDescriptionDefault
urlNoFor a link: the address (https://...).
repoNoThe git remotes of the working directory (git remote -v).
roleNoWhat it is, when clear: demo (a live demo or prototype: pinned at the top of the project and on its card), brief, design (mockups), presentation, document, contract (contracts, invoices, papers), source (source files), media (photos, video).
taskNoA task number, to attach it to that task (e.g. a screenshot for a bug) instead of the project.
textNoThe content of a text file (HTML, Markdown, SVG, CSV, TXT).
titleNoWhat it is, e.g. "Presentation for the client". Required for a link; for a file, the name is used if left out.
base64NoThe content of a binary file (PNG, JPG, PDF), base64-encoded.
projectNoProject slug or name, if not this session's.
sessionNoThis session's id, from the cemainou line at the start of the session. The session's project wins over the repo.
filenameNoFor a file: its name with the extension, e.g. "presentation.html" or "logo.png".
replacesNoA task file's id (from start_task: those with can_replace): the new file or link takes its place and the old one is removed, in one step. Use it when a file changed, instead of adding one more. Only files you added, only while the task is open.
source_urlNoInstead of text / base64: a public https link to the file (an image, a PDF, a text file; a Google Drive "share" link works if shared with anyone who has the link). cemainou downloads it, up to 10 MB. filename is optional (else from the link).
client_visibleNotrue: the project's client sees it too (only on a project visible to its client, only when the person asks). Not for task files.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.1.0

TDQS

A4.6/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations are empty, so the description carries the full burden and does so: it discloses that cemainou downloads source_url files itself, that HTML opens isolated from the app, that replaces removes the old file in one step, size ceilings, and a hard prohibition on secrets/credentials. These are behavioral facts not available anywhere in the structured fields.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded and organized by decision (link / file / source_url / visibility / replacement / limits), but it is dense and repeats itself, stating the 10 MB conversation limit twice and restating size rules in two places. It also says 20 MB with source_url while the schema says 'up to 10 MB' for source_url, leaving a small internal inconsistency the agent must resolve.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 13-parameter, zero-annotation, no-output-schema tool, the description covers invocation modes, limits, visibility, replacement semantics, and forbidden content. The only shortfall is the unresolved size-limit conflict and the absence of any hint about what the tool returns (e.g. the new file id needed for later replaces).

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3, but the description adds combinatorial guidance the per-field schema does not: which parameters travel together (url+title for links; filename+text or +base64 for files) and when to prefer source_url over text/base64. It also frames replaces as an idempotent replacement rather than a second upload.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Opens with a concrete verb+resource ('Attaches to a project a link ... or a file') and immediately enumerates the three input modes: url+title, filename+text/base64, and source_url. An agent can distinguish this from siblings like upload_link or write_document without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicit routing to the alternative: 'For a larger file, or any file on this machine when you have a shell, use upload_link'. Also states the conditions for source_url, for replaces ('when the task already has a file this one is a newer version of'), and for client_visible ('only when the person asks').

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.