# mksglu/context-mode

Context window optimization for AI coding agents. Sandboxes tool output (98% reduction), persists session memory, and   enforces routing across 17 platforms via MCP + hooks.

Repository: https://github.com/mksglu/context-mode
Canonical: https://ross.abutalabs.com/products/context-mode
Homepage: https://context-mode.com
Language: TypeScript
License: NOASSERTION
License Family: other
Topics: claude, claude-code, claude-code-plugins, mcp, skills, codex, copilot, opencode, antigravity, kiro, openclaw, claude-code-hooks, claude-code-skill, codex-cli, cursor-plugin, mcp-server, mcp-tools, pi-agent, zed-extension, context-mode
Last push: 2026-08-26T12:06:49+00:00

## Health v2 (maintenance only)
Score: 74/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 78, longevity 13
- inputs: {"age_days": 191, "days_push": 7, "days_rel": 65, "gap_med": 0.0, "n_releases_24m": 195}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 20167, forks 1461 (observed 2026-08-28T04:11:29.771493+00:00)

## What it is
An open-source MCP plugin that optimizes AI coding agent context windows by intercepting large tool outputs, storing them in a local FTS5 index, and letting the agent search instead of re-sending raw data. It also persists session memory and enforces routing across 17 AI platforms including Claude Code, Cursor, Copilot, and Codex.

## Use cases
- reduce token usage in claude code
- stop ai agent from burning context window
- sandbox tool output for llm agents
- persist session memory across ai coding sessions
- make copilot or cursor use fewer tokens
- search tool output instead of resending it
- optimize mcp server context
- cut llm api costs for coding agents

## When to choose
- you use AI coding agents heavily and hit context window limits
- you want to reduce LLM input token costs from repeated tool output
- you need session memory persistence across agent resets
- you work across multiple AI coding tools and want consistent context handling

## When to avoid
- you need a permissively licensed dependency (it uses Elastic License 2.0)
- you want a fully managed cloud analytics platform rather than a local plugin
- your agent workflows involve minimal tool output and context is not a bottleneck

## Facets
- artifact type: plugin
- maturity: active
- function: mcp, caching, search-engine, developer-tools, llm-inference
- domain: developer-tools, large-language-models
- platform: cli, cross-platform
- tags: context-window-optimization, mcp-server, claude-code, ai-coding-agents, token-savings, session-memory, tool-output-sandboxing, fts5, cursor, copilot, ai-agents, nodejs, docker

## Member repositories
- mksglu/context-mode (main) score 74

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:29.771493+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:59:19.615220+00:00, confidence not recorded.
  - readme: https://github.com/mksglu/context-mode (fetched 2026-08-28T04:11:29.771493+00:00, sha 164daee9bd0a)
  - homepage: https://context-mode.com (fetched 2026-08-29T07:58:08.747410+00:00, sha 344f0b586570)
  - site_page: https://context-mode.com/context-saving (fetched 2026-08-29T07:58:08.756924+00:00, sha a7fa99d0b6d2)
  - site_page: https://context-mode.com/insight (fetched 2026-08-29T07:58:08.759388+00:00, sha 0184e1a19595)
- Data as of 2026-08-30T08:39:29.467469+00:00.