multi-ai-frameworklisted

# Multi-AI framework Full specification lives in this skill's companion documents — read the one relevant to the question at hand rather than re-deriving it: - `SKILL_BRIEF.md` — what this is, why it exists, one-paragraph mental model - `REQUIREMENTS.md` — what "using the framework" actually obligates you to do - `BEHAVIOR_SPEC.md` — the hardcore-boss posture: inspect, verify, don't trust - `WORKFLOW.md` — the concrete step sequence for a delegated task - `TOOLS_AND_MCP.md` — each CLI's real invocation path, known quirks, when to use which - `PROMPT_ARCHITECTURE.md` — how to write a delegation prompt that gets real work back - `CONTEXT_MODEL.md` — what state carries across delegated calls, what doesn't - `QUALITY_GATES.md` — the checks a delegated result must pass before you act on it - `FAILURE_AND_RECOVERY.md` — known failure modes (hangs, stale claims, hallucination) and the fix for each - `TEST_SUITE.md` — how this skill itself was validated, and how to re-validate it - `EVOLUTION_PLAN.md` — what's deliberately out of scope for v1, and the trigger to revisit ## Standing defaults — apply automatically, every invocation, without being re-stated The user has corrected this multiple times (2026-07-29/30) after invocations that under-delegated or went shallow. These are not optional extras to remember only when explicitly asked for — they are the default mode every time this skill is invoked: 1. **Go deep by default — literally, every time.** The user's own standing phras