Context Mode
Sandboxes large tool outputs before they reach the model and keeps a persistent session memory, aiming to reduce context consumption in long agent runs.
What it does
Sandboxes large tool outputs before they reach the model and keeps a persistent session memory, aiming to reduce context consumption in long agent runs. The vendor claims a 98% reduction; read that as a ceiling on a good day, not a baseline. Matters most when an agent keeps re-reading the same files.
Install
npx -y context-mode
Client configuration
Paste this into your MCP client config (Claude Desktop, Cursor, or any client that reads mcpServers), then restart the client.
{
"mcpServers": {
"context-mode": {
"command": "npx",
"args": [
"-y",
"context-mode"
]
}
}
}
When to use it
Reach for this when the task needs context inside the agent loop. Skip it if a plain shell command or a one-off script is enough — every server you add costs context and adds a failure mode.
Other TypeScript servers
- Filesystem — Gives an agent scoped read, write, and search access to directories you explicitly allow
- Memory — A persistent knowledge graph stored in a local JSON file, exposed through entities, relations, and observations
- Sequential Thinking — Gives the model a scratchpad for structured, revisable reasoning: numbered thoughts that can branch, revise, and conclude