graphiq --help shows every flag. The full command guide for agents is in the skill and the AGENTS.md template .
Command
What it does
index <path>
Index a project (incremental by content hash)
search <query>
Ranked symbol search (name, natural language, file path, or error message)
context <symbol>
Read a symbol's source + structural neighborhood
blast <symbol>
Trace change impact (forward/backward radius)
impact --project <p>
Analyze current git changes for affected symbols
briefing [--compact]
Get oriented: architecture, subsystems, public API, hubs
interrogate
Ask structural questions (entry points, error boundaries, coupling)
topology <symbol>
Map the local graph neighborhood
constants [query]
Find shared numeric/string constants
dead-code
Find unreachable symbols
status / doctor / upgrade-index
Index health, diagnosis, rebuild artifacts
clear
Delete the index, leave a fresh empty database
Lifecycle & harness commands
Command
What it does
setup
Wire graphiq into agent harnesses
sync [--apply]
Verify harness attach + write the registry; --apply re-configures missing ones
discover
Scan the system for installed agent harnesses
git
Current project scope: branch, working tree, recent commits
projects [--forget <path|all>]
List tracked projects + their index health
--db <path> — use a specific index database (default .graphiq/graphiq.db)
--top <n> / --depth <n> — result count / traversal depth
--json — machine-readable output (for discover, git, projects, impact)
--yes — skip confirmation prompts (clear)