By oxagenai
Ontology layer for AI agents. Typed, queryable knowledge graph exposed to Claude Code as MCP-native tools — every code edit, research question, incident triage, or memory lookup gets grounded in deterministic graph context instead of guessing.
Show full context for a security alert — description, affected nodes, recommended fix, and links to upstream advisories.
Ask a natural-language question about the workspace ontology and get a direct answer backed by Cypher — faster than `/oxagen:search` for factual lookups.
Enriched blame for a file — line-level authorship plus the PRs and issues that touched each block.
List functions and methods that a symbol calls (its outgoing CALLS edges) ordered by PageRank.
List callers of a function or method — direct callsites in the workspace ontology, ordered by PageRank.
Save findings to the Oxagen knowledge graph after every meaningful step. Use whenever you make a non-trivial decision, learn something new about the codebase, or finish a task — so the next agent (or the human) can find it via memory_search.
Full catalog of Oxagen MCP tools — code-graph (discovery, traversal, history, tests, CI), ontology typed-traversal, security, agent coordination, memory, and general-purpose. Load on demand from the using-oxagen gate skill when a specific tool surface is needed.
Always-on routing policy for the Oxagen plugin. Use whenever the user asks a substantive question about the workspace, requests non-trivial changes to artifacts the workspace tracks (code, docs, runbooks, contracts, tickets, incidents), wants memory recall, or coordinates with other agents on the same workspace. Pairs with oxagen-tool-reference (full catalog) and oxagen-dogfooding (write-back).
External network access
Connects to servers outside your machine
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Ontology layer for AI agents. Typed, queryable knowledge graph exposed to Claude Code as MCP-native tools so every answer, edit, triage, or recall is grounded in deterministic graph context instead of guessing.
oxagenai/claude-code-pluginInstall once. Get:
/oxagen:* covering discovery, traversal,
history, tests, CI, security, memory, incidents, and multi-agent
coordination./clear and /compact.# Add the Oxagen marketplace
/plugin marketplace add oxagenai/claude-code-plugin
# Install the plugin
/plugin install oxagen@oxagen
That's it — no manual symlinks, no ~/.claude/plugins editing. The
manifest, MCP server config, hooks, commands, and skills all come from the
marketplace tree.
If you've cloned this repo and want to dogfood your changes:
ln -s "$(pwd)" ~/.claude/plugins/oxagen
Then restart Claude Code.
/oxagen:setup
The command opens https://app.oxagen.ai/setup/claude-code in your
browser, walks you through workspace + GitHub repo selection, hands back a
one-line install string, writes ~/.claude/oxagen.local.json, polls until
the first ingest is ready, and verifies the MCP tool list is live.
Add to your shell profile so the MCP token survives restarts:
export OXAGEN_MCP_TOKEN="$(jq -r .token ~/.claude/oxagen.local.json)"
Verify with:
/mcp
You should see oxagen listed with its full tool set.
When the Oxagen plugin is enabled, on every session the SessionStart
hook injects an <EXTREMELY_IMPORTANT> preamble that tells the agent:
ontology.*,
code.*, incident.*, security.*).memory.recall (or memory.context_inject) on the intent. If a
procedure or episode matches, prefer it over rediscovery. After
meaningful work, persist what you learned.agent.list_active_claims and agent.claim_work
when other sessions may be touching the same scope.The full routing policy lives in the oxagen:using-oxagen skill, which
is loaded into the system prompt by the hook on startup, clear, and
compact. The complete tool catalog and write-back conventions live in
the oxagen:oxagen-tool-reference and oxagen:dogfooding skills, which
are loaded on demand.
The plugin's value is the typed graph + memory layer. Coding is the most obvious surface. It is not the only one. Anything Oxagen ingests becomes queryable and groundable.
/oxagen:impact <symbol> returns
callers, dependents, PageRank, and reach paths before you ship a
signature change. Stops the "I didn't realize ten things called this"
class of regression./oxagen:repo-overview, /oxagen:module-tree,
/oxagen:who-knows <symbol>, /oxagen:expert <login> — the new hire
gets a cartography of the system and its owners in five minutes
instead of three weeks./oxagen:ask "list nodes with zero callers and no EXPORTS edge" returns provably dead symbols. Safe to delete./oxagen:ask "show cycles in the call graph scoped to packages/oxagen" surfaces SCCs you can't see from grep.npx claudepluginhub oxagenai/claude-code-plugin --plugin oxagenHarness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.