Graph setup and settings

Index a project

From any project:

ctx graph index .
ctx graph stats
ctx graph search authenticate

Indexing and updates are explicit. Reads use the saved snapshot and do not refresh source files. After changing code, update only what changed:

ctx graph update

ctx graph saves the graph at .graf/index.db and discovers the nearest ancestor database. ctx graph --db PATH ... selects one explicitly. Use an exact node ID when a symbol name is ambiguous. Inspect generation, coverage, unresolved references and truncation before judging completeness. Graph reachability describes potential impact, not proof of runtime behavior or agent authorship.

Search scope

ctx search "retry policy"
ctx search --scope graph authenticate
ctx search --scope all authenticate

The default search scope is history. graph searches indexed code and document relationships. all returns separately labeled history and graph results from retained snapshots. Scores from the two engines are not a common ranking scale. Use ctx graph --help for callers, paths, filters, reports, exports, and the full graph command set.

Use with an agent

ctx integrations install skill
ctx integrations install mcp --agent codex

Optional project graph tool hooks use ctx graph install --platform gemini --project PATH (or claude or codebuddy). Remove them with ctx graph uninstall and the same selections. Git refresh hooks use ctx graph hook install --project PATH.

Existing .graf/index.db databases work with ctx graph; no conversion is required. Existing Graf integrations are not implicitly removed or rewritten.

For full source and platform details, see the ctx graph guide.