Repository navigation
Proposal: Add source path and untrusted flag to read results #5151
Description
Which server?
filesystem (@modelcontextprotocol/server-filesystem)
Protocol era
Both, or not era-specific
MCP client
No response
The problem
Read tools currently return file contents as plain text, and structuredContent holds only that same text, so hosts get no provenance or untrusted field they can act on.
Solution you have in mind (optional)
Return the source path and an untrusted-content flag in structuredContent for every read tool (index.ts#L208-L211).
Came up while reviewing this project for SafetyCard: https://safetycard.dev/github.com/modelcontextprotocol/servers/src/filesystem
Which MCP feature does it demonstrate? (optional)
No response
Alternatives or workarounds you have tried
No response
Already built it locally?
No response
- I searched existing issues and this is not a duplicate.
- This is a request for this repository or one of its servers, not a new server, the MCP specification or an SDK.
Activity
[-]Proposal: Add source path and untrusted flag to read results (SafetyCard review)[/-][+]Proposal: Add source path and untrusted flag to read results[/+]cliffhall commentedon Oct 19, 2026
Closing: out of scope for the reference servers.
Thank you for the suggestion, and apologies for the long wait for a response.
These servers are reference implementations: they exist to show how each part of MCP is used, not to be general-purpose products, so new features are taken selectively (
CONTRIBUTING.md). An untrusted-content flag is not defined by the MCP specification, so adding an ad hoc provenance field to a reference server's results would not demonstrate a protocol feature. This is not a judgment on the idea or on you; a server in the MCP Registry may well offer it, and anyone is welcome to build on these servers to do so. This is happening as part of the backlog triage in #4875.