A Claude Code plugin that turns Claude into a graph-reasoning engineering agent using Ix Memory as its structured memory backend. Claude + Ix = reasoning engine + persistent code knowledge graph. Skills are cognitive abstractions (not CLI wrappers) that minimize token usage and maximize accuracy.
Installation
/plugin marketplace add ix-infrastructure/ix-claude-plugin
/plugin install ix-memoryRestart Claude Code after installing. Requires Ix Memory installed and running (ix status returns ok), jq, and rg in PATH. Ix Pro is optional — all skills and hooks work with basic Ix. Pro adds session briefing injection (goals, bugs, decisions) via the ix-briefing.sh hook.
Skills
High-level cognitive skills — each one infers intent, orchestrates multiple graph queries, and synthesizes output. None are CLI aliases.
/ix-understand [target]Build a mental model of a system or the whole repoGraph only — no source reads
/ix-investigate <symbol>Deep dive: what it is, how it connects, execution pathGraph first, one symbol read max
/ix-impact <target>Change risk: blast radius, affected systems, test targetsDepth scales with risk level
/ix-plan <targets...>Risk-ordered implementation plan for a set of changesParallel impact, finds shared dependents
/ix-debug <symptom>Root cause analysis from symptom to candidatesTargeted reads at suspects only
/ix-architecture [scope]Design health: coupling, smells, hotspotsGraph only — never reads source
/ix-docs [options]Generate documentation from graph structureSupports --full, --style, --split
Agents
Autonomous multi-step agents for complex tasks.
ix-explorerGeneral-purpose graph exploration, open-ended questions
ix-system-explorerFull architectural model of a codebase or region
ix-bug-investigatorAutonomous investigation from symptom to root cause candidates
ix-safe-refactor-plannerBlast radius + safe change sequencing for refactors
ix-architecture-auditorFull structural health report with ranked improvements
Automatic Hooks
User sends any promptInjects session briefing (goals, bugs, decisions) once per 10 min — requires Ix Pro
Claude runs Grep or GlobFront-runs with ix text + ix locate/ix inventory
Claude runs ReadFront-runs with ix inventory + ix overview for the file
Claude runs Bash with grep/rgExtracts pattern, front-runs with ix text + ix locate
Claude edits a file (pre)Runs ix impact before the edit
Claude edits a file (post)Runs ix map <file> to update the graph (async)
Claude finishes respondingRuns ix map to refresh the full graph (async)
All hooks bail silently if ix is not in PATH or the backend is unreachable. Uninstall with /plugin uninstall ix-memory. See the ix-claude-plugin repo for full details.
03 — Codex & Agent Integration Ix ships an AGENTS.md file that any agent (OpenAI Codex, Claude, Cursor, etc.) can read to use the ix CLI effectively. Drop it in your repo root and the agent will use Ix automatically.
Setup
cp AGENTS.md your-project/AGENTS.mdWhat the agent gets
JSON modeAll commands support --format json for structured output
Command routingHigh-level workflows (overview, impact, rank) before low-level primitives
Chaining patternsStart broad with overview/impact, drill into callers/callees
Diagnostic codesStructured error handling (no_edges, ambiguous_resolution, etc.)
Planning & decisionsGoals, plans, tasks, bugs, and decisions with linked workflows
Architecture analysisDetect code smells, score subsystem health, find hotspots
Recommended agent loop
ix briefing --format json # resume session context
ix overview <target> --format json # understand a component
ix impact <target> --format json # check blast radius
ix rank / ix inventory # find hotspots or list entities
ix plan next <id> --with-workflow # get next task
ix tasks --status pending # see all pending tasks
Works with any agent that reads markdown instruction files. See AGENTS.md for the full reference.
The 6 commands that cover 90% of daily use. See Every Command for the full reference.
ix mapParse files, build knowledge graph, detect architecture
ix search <term> --kind classFind entities by name with ranked relevance
ix explain <symbol>Role, importance, callers, dependents — no source reading
ix impact <target>Blast radius — what breaks if you change this?
ix rank --by dependents --kind class --top 10Find the most structurally important entities
ix read <symbol>Read source code by name or file:line range