Research Context

Homingo sits in an emerging category of tooling for large AI skill ecosystems.

Two recent papers are especially relevant:

• "Organizing, Orchestrating, and Benchmarking Agent Skills at Ecosystem Scale"
• "SkillOrchestra: Learning to Route Agents via Skill Transfer"

What Those Papers Focus On

Both papers study what happens when an agent system has too many skills for flat, implicit routing to stay reliable.

They emphasize:

• Structured skill representations instead of a flat list of descriptions
• Retrieval or routing at ecosystem scale, not just one-off tool calls
• Benchmarks and empirical evaluation to measure whether the right capability is selected
• Runtime orchestration across many candidate skills, agents, or models

Where Homingo Fits

Homingo solves a different but closely related problem: skill fleet governance.

Homingo is strongest before or alongside runtime orchestration:

• homingo scan finds overlap, duplicates, and overloaded skills with no API calls
• homingo audit measures routing ambiguity with adversarial prompts
• homingo lint suggests rewrites, merge candidates, and shard plans

That means Homingo is not primarily a runtime agent orchestrator. It is the layer that helps you verify whether your skill fleet is understandable, distinguishable, and maintainable before routing failures accumulate in production.

How Homingo Differs

The papers and Homingo are complementary, but not interchangeable.

The papers focus more on:

• runtime retrieval and orchestration
• learned routing policies or reusable skill handbooks
• benchmarked downstream task success

Homingo focuses more on:

• routing drift caused by overlapping descriptions
• pairwise ambiguity detection and fleet hygiene
• practical remediation: rewrites, merge recommendations, and sharding

Why This Matters

If your runtime system is getting better at orchestration, that does not remove the need for Homingo.

Better runtime routing still benefits from:

• cleaner skill boundaries
• fewer redundant descriptions
• less overloaded scope
• a clearer capability map for humans and machines

Homingo improves the quality of the skill fleet itself, which makes downstream routing and orchestration more reliable.

How These Papers Influence Homingo

These papers help motivate several extensions to Homingo's roadmap:

• homingo map to produce a capability map and handbook-style artifact for the fleet
• homingo audit --mode fleet to complement pairwise ambiguity testing with full-manifest routing evaluation
• stronger benchmark and profile-based evaluation over time

Further Reading

• Routing Drift
• Shadow Router
• homingo audit