What problem does it solve?
It solves inefficient and error-prone multi-agent coordination by enforcing parallel dispatch safety, clear task ownership, and verification-before-completion so agents don’t race, drift, or claim incorrect outcomes.
Core Features & Use Cases
- Parallel swarm dispatch for reviews and pipelines: coordinate multiple agents for independent analysis or multi-stage workflow execution with explicit fan-out rules.
- One-owner-per-file safety with pre-dispatch checks: prevent overlapping edits by requiring task prompts that declare owned files and by adding an operational intersection gate.
- Two-stage subagent verification: verify spec compliance first, then assess correctness and quality to avoid “polished answers to the wrong problem.”
- BLOCKED triage and resilience patterns: classify blockers by missing context, reasoning limits, task size, or spec errors, and respond with re-dispatch, escalation, or decomposition rather than blind retries.
Quick Start
Use the ia-orchestrating-swarms skill to run a multi-agent review that parallelizes independent module analysis while enforcing spec compliance checks and safe file ownership before committing results.
