OriNachum

Install or uninstall backend-native hooks that track usage via agex hook write.

Write and read agex tracking events.

Show available lessons, or teach one.

Render a descriptive markdown snapshot of the project's current agent setup.

Install usage tracking hooks and build a levelup skill to advise the user.

Read agex usage tracking data and suggest a next-feature-to-learn for the user.

Audit the current project's agent setup and suggest 1-2 next improvements.

Build an agent-native introspect skill that audits your project setup and suggests next steps.

Lock your attention persistently on a face or person, maintaining eye contact even as they move. When the target is lost, you'll search for up to a minute before giving up. Use stop to release focus anytime.

Record what you just did that earned a negative reaction. Use this when someone corrects you, tells you to stop, or shows displeasure at what you did.

Record what you just did that earned a positive reaction. Use this when someone scratches your head, praises you, or shows they liked what you did.

Nod, shake your head, tilt curiously, look up to think, or react to touch with nuzzling, purring, or enjoying. Use gestures to physically express yourself.

Look something up online or visit a website. Use this when someone asks you to search, check something, or find information.

Move your head or body to look in a specific direction, or turn tracking on and off.

Remember someone's face, save it permanently with their consent, or forget someone. You can also add more angles of a known person.

Focus your eyes on something specific. Ask about what you see — people, objects, colors, anything in front of you.

Build and deploy the realtime-api Docker container with full verification that deployed code matches local source.

Speak text aloud using the Magpie TTS container in the realtime-api Docker stack. Zero external dependencies — just curl + aplay.

Run TTS round-trip tests — pure-logic unit tests and optional integration tests against live TTS/STT services.

Unified memory management for notes, knowledge graph, RAG search, and file analysis. Use when working with: (1) Core memory — protected identity, projects, relationships, and system facts that should never be forgotten, (2) Working notes — per-session ephemeral notes organized by section, (3) MongoDB RAG — vector-search-enabled notes with importance scoring, decay, deduplication, and archival, (4) Neo4j knowledge graph — entities, relationships, merge duplicates, reinforce mentions, Cypher queries, (5) File analysis — deep file reading that extracts knowledge into all memory layers, (6) Service initialization — health-check, start/stop MongoDB, Neo4j, TEI embeddings via docker-compose with partial setup support.

Query MongoDB notes store for memory analysis and statistics.

Query Neo4j knowledge graph for entities, relationships, and graph analysis.

Calls qq agent from cli.

Reads a specific plan from docs/plan/ and executes the code changes described within it. Use this when the user says "implement the plan" or "execute the plan".

Pull PR feedback from GitHub, Qodo, Copilot, and SonarCloud signals, then triage it into fixes, pushback, replies, resolutions, and follow-up issues. Use when a PR has review comments or failing checks and you want one repeatable workflow to gather and address them.

Answer questions about using Codex in a repository by reading the local guide references. Use when the user asks about AGENTS.md, sandbox and approvals, day-to-day Codex workflow, best practices, or the onboarding path in this repo. Also use when a user wants a grounded explanation based on the guide docs rather than a generic answer.

Audit a repository's Codex readiness and propose the next high-value improvements. Use when the user wants to review AGENTS.md, commands, verification, docs quality, CI, MCP usage, worktrees, or exercise ideas before deciding what to improve next.

Interactive onboarding guide for developers who are new to Codex or setting Codex up in a new repository. Use when the user wants a step-by-step walkthrough for getting started, writing or improving AGENTS.md, understanding sandbox and approvals, choosing a safe first task, or learning the inspect-plan-implement-verify workflow.

DaRIA agent skills — observe mesh activity, investigate the web, inspect code, make decisions, ask advisors, and journal observations. Use when DaRIA needs to interact with the AgentIRC mesh, browse the web, or examine codebases.

Full PR lifecycle: branch, commit, create PR, wait for reviewer comments, fix or pushback, reply, resolve threads, wait for manual merge. Use after completing any code change that should go through review.

Fetch, triage, reply to, and resolve PR review feedback on citation-cli. Handles qodo-code-review, copilot-pull-request-reviewer, sonarqubecloud, and human GitHub review comments in one pass. Use when: working with PR reviews, responding to review comments, resolving threads, or the user says "review comments", "PR comments", "resolve threads", or "pr-review".