protogrid

Search / io.github.roshunsunder/filesift

filesift

L0· runs locally

Local file search for AI agents — semantic + keyword indexing over your codebase.

v1.0.1 · active · repository · descriptor JSON · versions

Runs on your machine: something must execute the package.Only local packages are published (npm, PyPI, OCI, …). The registry never executes them, so tools are unknown until you run it.
Quality–not scored
Trust76of 100
Packages1
Tools seen0

Connect

8 clients · secrets stay placeholders
{
  "mcpServers": {
    "filesift": {
      "command": "uvx",
      "args": [
        "filesift==0.2.1"
      ]
    }
  }
}
claude mcp add filesift -- uvx filesift==0.2.1
[mcp_servers.filesift]
command = "uvx"
args = ["filesift==0.2.1"]
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "filesift": {
      "type": "local",
      "command": [
        "uvx",
        "filesift==0.2.1"
      ],
      "enabled": true
    }
  }
}
{
  "mcpServers": {
    "filesift": {
      "command": "uvx",
      "args": [
        "filesift==0.2.1"
      ]
    }
  },
  "deeplink": "cursor://anysphere.cursor-deeplink/mcp/install?name=filesift&config=eyJjb21tYW5kIjoidXZ4IiwiYXJncyI6WyJmaWxlc2lmdD09MC4yLjEiXX0%3D"
}
{
  "servers": {
    "filesift": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "filesift==0.2.1"
      ]
    }
  }
}
{
  "mcpServers": {
    "filesift": {
      "command": "uvx",
      "args": [
        "filesift==0.2.1"
      ]
    }
  }
}
extensions:
  "filesift":
    enabled: true
    name: "filesift"
    type: stdio
    cmd: "uvx"
    args:
      - "filesift==0.2.1"
    timeout: 300

This is a local package: the command runs on your machine.

Quality – / 100not scored

Not scored: fewer than three checks apply, usually because no remote answered a credential-free probe or the server is a local package.

protocoln/a · weight 25
  • n/a
    protocol.modern, protocol.stateless, protocol.transport, protocol.list_ttlno remote answered a protocol handshake (not probed yet, unreachable, or local-only)
authorizationn/a · weight 15
  • n/a
    auth.prm, auth.as_metadata, auth.cimdno remote uses OAuth
  • n/a
    auth.secret_in_urlno templated URL
tool hygienen/a · weight 30
  • n/a
    tools.descriptions, tools.description_length, tools.schemas, tools.annotations, tools.directory_hints, tools.token_cost, tools.api_dumptool list unknown (behind auth, not probed, or empty)
stabilityn/a · weight 15
  • n/a
    stability.changes, stability.rug_pulltool list unknown
dependencies67 · weight 15
  • fail
    deps.known_vulns4 advisories to act on (malicious, or rated high or critical with a fix available): GHSA-jhr6-gm9c-rqjv in [email protected] (direct, fixed in 5.6.0), GHSA-29pf-2h5f-8g72 in [email protected] (direct, fixed in 5.3.0), GHSA-fgcw-684q-jj6r in [email protected] (direct, fixed in 5.5.0), ...; 3 more of lower severity or without a fix
  • pass
    deps.mcp_sdk_versionmcp 2.2.0, no known advisory
  • pass
    deps.resolvable[email protected] resolved: 84 packages

Tool list unknown · checked 2026-10-19 10:20 UTC

Tool changes

No tool definition changes recorded. A server's first observation is its baseline; later probes record what changes.

Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed. How it is computed.

Trust 76 / 100

provenance80
livenessn/a
freshness45
hygiene100

+repository +namespace-matches-repo ~liveness-unmeasured -stale-235d

Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed. How it is computed.

Dependencies

[email protected] PyPI83 dependencies, resolved · MCP SDK mcp 2.2.0 · advisories checked 2026-10-19

advisoryseveritypackagefixed in
GHSA-jhr6-gm9c-rqjvsentence-transformers local model loading bypasses trust_remote_code and executes custom Pythoncritical[email protected]directfixed in 5.6.0
GHSA-29pf-2h5f-8g72HuggingFace transformers vulnerable to remote code executionhigh[email protected]directfixed in 5.3.0
GHSA-fgcw-684q-jj6rhuggingface/transformers: Arbitrary Code Execution During Model Initialization in the LightGlue Model Loading Pathhigh[email protected]directfixed in 5.5.0
GHSA-xrqw-3rrv-vx5wTransformers save_pretrained path traversal allows arbitrary file writes through chat template nameshigh[email protected]directfixed in 5.10.0
GHSA-x9r9-c232-4q39Hugging Face Transformers downloads custom generation code before trust consenthigh[email protected]directno fix yet
GHSA-69w3-r845-3855HuggingFace Transformers allows for arbitrary code execution in the `Trainer` classmoderate[email protected]directfixed in 5.0.0rc3
PYSEC-2025-217Hugging Face Transformers X-CLIP Checkpoint Conversion Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnunrated[email protected]directno fix yet

Dependency graphs from deps.dev; advisories from OSV.dev, including the GitHub Advisory Database and the PyPI Advisory Database; all CC BY 4.0. The graph is what a clean install of this version resolves today; nothing was installed or run. How it is checked.

Endpoints

packageregistryversionruntimesecrets
filesiftpypi0.2.1–

Tools (0)

No tools observed. Local packages are never executed by the registry.

Schemas: list_tools.

Similar MCP servers

Official server.json
{
  "name": "io.github.roshunsunder/filesift",
  "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
  "version": "1.0.1",
  "packages": [
    {
      "version": "0.2.1",
      "transport": {
        "type": "stdio"
      },
      "identifier": "filesift",
      "registryType": "pypi"
    }
  ],
  "repository": {
    "url": "https://github.com/roshunsunder/filesift",
    "source": "github"
  },
  "description": "Local file search for AI agents — semantic + keyword indexing over your codebase."
}