Embedding, search, novelty detection, and clustering for knowledge-dense text corpora. Optimized for claims, research findings, notes, and annotations.
This repository also configures its own agents. See what limbic tells them →
Skill Codex
Turn a long, ambiguous voice dump such as 'research this' or 'improve this' into a bounded direction card grounded in nearby evidence. Use when the user wants an agent to drive an open-ended research, product, design, or coding effort; choose the right first move; avoid an expensive rabbit hole; or coordinate Codex…
Skill Claude CodeCodex
Keep a long-running coordinating thread from becoming the bill, and keep its destructive commands from becoming the incident. Use when running a multi-step or multi-agent project from one thread — deciding when to hand off, how to wait for a long job, when to run gates, or before any rm -rf / git checkout -- / reset /…
Skill Claude CodeCodex
Apply a codebook, extract structured records, or classify across N documents without spending agent tokens on the work itself. Use when the task is 'code/extract/classify/enrich these documents' at a scale above a handful — building a stateless packet runner with deterministic candidate retrieval, fixed slot enums…
Skill Claude CodeCodex
Brief a subagent so it does not pay a 50K-token entrance fee to rediscover the project. Use when delegating to workers, reviewers or researchers — writing a WORKER.md, choosing a model tier for a delegated role, setting a tool-call budget, or deciding whether to delegate at all. Triggers on 'spawn a subagent'…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: