compress
Compress a memory/context file into terse form. Saves ~40-50% tokens per session. Usage: /compress <filepath>
Switch IJFW mode. Usage: /mode smart|fast|deep|manual|brutal
$ npx -y skills add FerroxLabs/ijfw --agent claude-codeHow it fires
How this command gets triggered: by you, by Claude, or both.
/modeContext preview
What this command does when you run it.
Switch IJFW mode. Usage: /mode smart|fast|deep|manual|brutal
name: mode description: "Switch IJFW mode. Usage: /mode smart|fast|deep|manual|brutal"
Switch the active IJFW mode. Acknowledge with one line.
**smart** (default): Auto-routes models, effort, and verbosity by task type. Routine → scout/builder, terse. Complex → architect, thorough. Teaching → normal verbosity.
**fast**: Maximum efficiency. Cheapest viable model. Low effort. Ultra-terse. Aggressive compaction. For blasting through known tasks.
**deep**: Maximum quality. Best model. High/max effort. Self-verification on. Plan-then-execute enforced. Normal verbosity. For architecture, security, complex debug.
**manual**: All IJFW automation off. User controls /model, /effort, verbosity. Output rules still active but overridable. For power users.
**brutal**: Caveman mode -- 1-sentence answers, code-only, no explanation unless user asks "why" or "explain". Reasoning/thinking preserved, output constrained. Sets IJFW_TERSE_ONLY=1. For blast-mode task execution.
Natural language triggers also work:
Mode persists until changed or session ends.
IJFW — It Just F*cking Works. Ferrox Labs' local-first infrastructure for AI coding agents: shared memory, smart routing, multi-AI cross-audits, disciplined workflow.
Repo: FerroxLabs/ijfw
Compress a memory/context file into terse form. Saves ~40-50% tokens per session. Usage: /compress <filepath>
Dream cycle -- promote patterns, prune stale, reconcile contradictions, optionally promote to global.
Second-model review. Picks an auditor (codex/gemini/opencode/aider/copilot), excludes the caller, writes a prompt to paste, reads the response back. Usage:…
Adversarial multi-angle critique. All three auditors fire in parallel (codex=technical, gemini=strategic, claude=ux). Counter-args ranked by rebuttal survival…
Two-phase multi-model research. Phase A fans codex+gemini in parallel (benchmarks + citations angles); Phase B synthesises via fresh Claude session. Usage:…
Run IJFW health check (files, MCP server, hooks, memory, caps, framing)