Plugin Claude Code
its-over-9k (o9k) — the connector between agent frameworks for Claude Code, Cursor, Codex, OpenCode and Hermes. Arbitrates which tool owns which concern, wires them into every host, and keeps the interplay efficient. Ships no features of its own.
Plugin Claude Code
Output compression: telegraphic response style with lite/full/ultra levels and automatic fallback to full prose for safety-critical content. Adapted from JuliusBrussee/caveman (MIT).
Skill Claude CodeCodex
Token-compressed output style — telegraphic fragments, zero filler, code byte-exact. Levels: lite/full/ultra. Use when the user enables caveman mode, asks for terse/compressed output, or the o9k doctrine is active. Auto-reverts to full prose for safety-critical content.
Plugin Claude Code
The o9k doctrine: token-efficiency rules of engagement and the arbitration layer that keeps efficiency plugins from colliding. Zero-effort: a SessionStart hook activates the doctrine automatically, checks dependencies for updates (npm o9k package + git marketplace channels; refreshes multi-CLI skills/hooks after…
Hook
Runs when a session starts, executing session-start.mjs and update-check.mjs via node (2 commands). From Bumblebiber/its-over-9k.
Skill Claude CodeCodex
Personalized o9k orientation for the human. Use when the user invokes /o9k-guide, asks how o9k works or what they need to do, right after first-time setup, or when the SessionStart hook flags an unresolved arbitration. Explains what runs automatically, the few one-time actions left, and on-demand commands (/o9k-init…
Skill Claude CodeCodex
Guided first-install and reconfiguration flow for o9k. Use when the user invokes /o9k-init, on a fresh o9k install, or when the user wants to set up, reconfigure, or extend their companion stack. Detects what's installed, interviews the user (bundle, git, conflicts), resolves collisions with migration, then executes…
Skill Claude CodeCodex
Token usage report for the current project from Claude Code session transcripts. Use when the user asks how many tokens were used/saved, invokes /o9k-stats, or wants to verify the o9k pillars are paying off.
Skill Claude CodeCodex
Check installed o9k pillars and companion frameworks for updates, and apply the safe ones. Use when the user invokes /o9k-update, asks whether anything is out of date, or acts on the SessionStart 'updates available' notice.
Skill Claude CodeCodex
Meta-skill for the o9k doctrine and arbitration — who owns output style, the repo map, dispatch, memory, the plan. Load at session start, on conflicts between efficiency skills, and whenever unsure Path A vs Path B — if a team-up roster exists, external CLI tmux workers require team-up + dispatch mailbox/watcher (not…
Plugin Claude Code
Persistent memory integration: session-start briefing, on-demand recall, save-before-compact. Prefers TIM (npm tim-cli) when present; falls back to hmem (npm hmem-mcp).
Skill Claude CodeCodex
Memory-MCP integration for o9k. Use at session start (load the project briefing), when the user references past work ('the bug we had', 'where did we leave off'), before compaction or /clear (flush state), and when deciding what deserves persisting. Works with TIM (npm tim-cli) or hmem (npm hmem-mcp); hooks pick TIM…
Plugin Claude Code
Discovery & onboarding: a GitHub Scout skill for finding new agent frameworks and slotting them into the o9k compatibility matrix, one-command companion bundle installs, and a bundle benchmark for measuring companion combos.
Skill Claude CodeCodex
Benchmark companion combos on a fixed task set to find the best-measured o9k bundle. Use when the user wants to test/compare companion combinations, validate the existing setup with numbers, or contribute benchmark results back to the o9k repo. Wraps benchmarks/run-bench.sh with sandbox isolation and an ablation plan.
Skill Claude CodeCodex
Install a conflict-free stack of third-party o9k companion frameworks in one command. MUST use when the user asks to set up companions, install the recommended stack, add memory/docs/methodology/symbol tools alongside o9k, or extend an existing bundle — never hand-install MCP servers without checking arbitration…
Skill Claude CodeCodex
GitHub Scout for agent frameworks. MUST run before installing an unknown companion, extending a bundle, refreshing the o9k compatibility matrix, or answering 'what's new/better than X' for Claude Code plugins, MCP servers, or agent frameworks. Repeatable recon loop: hunt, score, classify against using-o9k arbitration…
Plugin Claude Code
Multi-agent model registry: role-based model selection with fallback chains, usage-limit awareness, tmux worker dispatch, limit-triggered session handoff, mailbox runs + crash resume. Selection is deterministic code, not LLM reasoning.
Skill Claude CodeCodex
Refresh o9k-roster scores/prices from OpenRouter (Artificial Analysis indices + model catalog), then semiauto-apply chain updates when score rises and cost does not. Use weekly, after major model releases, or when the user asks to update the matrix. Includes hosted open-weight models for Hermes/OpenCode.
Skill Claude CodeCodex
Role-based model selection for multi-agent delegation. Use only when o9k-roster is installed and /.o9k/roster.json exists — before delegating to another model/CLI, on rate-limit errors (mark-limited), session-limit handoff, or cross-CLI mailbox runs (create/wait/answer/resume). Not for ordinary in-host search…
Plugin Claude Code
Context discipline: load structure before files, search before read, one canonical repo map per session.
Skill Claude CodeCodex
Context-loading discipline for codebases. Use BEFORE opening, reading, or exploring files — and for monorepo/broad greps (TODO sweeps, 'where is X'): search first, read narrow, one repo map per session. Pair with dispatch path A when the sweep would pollute main context. Prevents the 20k-token 'let me look around'…
MCP server Claude CodeCodexCursor +2
its-over-9k (o9k) — token-efficiency meta-framework for Claude Code, Cursor, Codex, OpenCode, and Hermes. Runs locally from the its-over-9k npm package.