Model Context Protocol moderate risk

LLM Context.py

llm-context provides intelligent context management for LLMs, enabling focused, task-specific project context through composable rules and MCP integration.

Connections & Capabilities

Connects To

GitHubTwitter

Capabilities

readwriteexec

Quickstart

Config

{
  "mcpServers": {
    "CyberChitta": {
      "command": "uvx",
      "args": ["--from", "llm-context", "lc-mcp"]
    }
  }
}

Exposed MCP Tools (3)

safe

lc_outlines

Generates code structure excerpts from the current rule's selection.

Read-only operation, only extracts code structure.

safe

lc_preview

Validates the rule effectiveness by showing file selection and size.

Read-only operation, only previews the rule's output.

moderate

lc_missing

Fetches specific files or implementations on demand.

Potentially exposes file contents, but requires explicit file paths.

Safety Assessment

llm-context offers a controlled way to provide context to LLMs, but the safety depends heavily on the rules defined. It is safe when used with carefully crafted rules that limit access to necessary files. It is risky if rules are overly broad or expose sensitive information.

Rules-based system limits scope of access.
MCP integration allows for controlled file access.
Preview functionality enables validation before context generation.
No direct internet access or external API calls by default.
Project-relative paths prevent accidental access to sensitive system files.

Can expose source code and project structure to LLMs.
Improperly configured rules could lead to excessive context exposure.
MCP integration relies on the security of the underlying OS and file system.
AI-assisted rule creation relies on the LLM's ability to generate secure rules.
Potential for information leakage if sensitive data is included in the project.