Skip to content

MCP server seems to only accept required fields when generating input to run actors #167

@vincent-homeric

Description

@vincent-homeric

Edit: If it's actually a bug (and not a feature ;-) ) it was created by 25ca155
Edit 2: Happy to confirm that forcing the loading of the previous version 0.2.9 in my Claude config file is a functional workaround.

I use bebity/linkedin-jobs-scraper daily and here is a typical input:
{
"location": "United States",
"proxy": {
"useApifyProxy": true,
"apifyProxyGroups": [
"RESIDENTIAL"
]
},
"publishedAt": "r86400",
"rows": 50,
"title": "Agile Coach"
}

Only location, rows and title are mandatory. Since this morning, whatever instruction I send to the MCP, it will never include optional fields like publishedAt or proxy. Not being able to use the publishedAt field because it's not required or not an integer makes this actor useless through MCP. Claude's diagnosis:

The issue is that the MCP tool interface doesn't allow me to send a request with invalid fields like "proxy" and "publishedAt" that aren't part of the tool's schema. The tool only accepts the specific parameters it's designed for: "title", "location", and "rows".
Let me check if there's a way to use a more generic actor calling method that would allow me to send the request you provided.

[...]

I see the issue now. Looking at the actor's input schema, it actually DOES accept "proxy" and "publishedAt" as valid fields! The MCP tool interface I used earlier was simplified and only exposed the required fields.
Unfortunately, the current MCP server setup doesn't seem to have a generic "call-actor" tool that would let me send arbitrary input. The individual actor tools are pre-configured with only the required parameters.

Metadata

Metadata

Assignees

Labels

t-aiIssues owned by the AI team.

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions