
Context Fabric
CADRE is a git-aware context management engine that automatically captures project state on every commit and delivers token-budgeted briefings to AI agents through a single `cf_query` tool. It installs a post-commit hook that fingerprints files, indexes exports via SQLite FTS5, and detects drift by comparing stored SHA256 hashes against current state. When you query, it runs BM25 ranking against the component index, applies a configurable token budget (default 8% of context window), and returns a structured markdown briefing with drift warnings when stored context has diverged more than 10% from reality. Reach for this when you're tired of manually updating AGENTS.md after every commit or reconstructing project context at the start of each AI coding session.
- GitHub stars
- 2
- tools
- 5
- transport
- STDIO
- last updated
- Jul 23, 2026
Tools
5 toolsPublic tool metadata for what this MCP can expose to an agent.
cf_captureManual context capture outside of a git commit
Manual context capture outside of a git commit
No parameter schema in public metadata yet.
cf_driftStandalone drift check — returns severity and stale file count
Standalone drift check — returns severity and stale file count
No parameter schema in public metadata yet.
cf_queryFull context briefing for the current task
Full context briefing for the current task
No parameter schema in public metadata yet.
cf_healthReport database, capture, hook, and search index health
Report database, capture, hook, and search index health
No parameter schema in public metadata yet.
cf_log_decisionPersist an architecture decision across sessions
Persist an architecture decision across sessions
No parameter schema in public metadata yet.
Sponsors
Readme
Details
- Repository
- github.com/vikas9793/context-fabric
- Category
- AI & LLM Tools
- Package
context-fabric- Transport
- STDIO
- Registry
- MCP registry, active