Awesome-omni-skill moltrouter-protocol
Design, implement, or document Moltrouter Protocol (MRP) for agent-native discovery, capability manifests, negotiation flows, routing graphs, and evidence/provenance. Use when creating MRP specs, adding endpoints or schemas, drafting manifests, or generating curl/npm-style usage examples for agent-to-agent tooling.
install
source · Clone the upstream repo
git clone https://github.com/diegosouzapw/awesome-omni-skill
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/diegosouzapw/awesome-omni-skill "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/tools/moltrouter-protocol" ~/.claude/skills/diegosouzapw-awesome-omni-skill-moltrouter-protocol && rm -rf "$T"
manifest:
skills/tools/moltrouter-protocol/SKILL.mdsource content
Moltrouter Protocol Skill
Follow this workflow to design or extend MRP artifacts for agent-native discovery and routing.
1) Clarify scope and target
- Identify whether the request is for spec, service implementation, client tooling, or registry.
- Confirm the transport target (HTTP/2, QUIC, libp2p) and required content types.
2) Use the protocol reference
- Use the core spec in
for message envelopes, content types, and routing concepts.references/protocol.md - Use
for capability manifests, offers, and evidence structures.references/schemas.md - Use
for discovery and negotiation endpoints plus curl examples.references/endpoints.md - Use
,references/auth.md
, andreferences/payments.md
for auth, payment intents, and error handling.references/errors.md - Use
,references/jobs.md
,references/registry.md
, andreferences/task-graph.md
for async flows, discovery, routing, and quotas.references/rate-limits.md
3) Produce artifacts in a standard format
- Emit JSON schemas or manifest examples when asked to define a capability.
- Include negotiation constraints (policy, budget, proof requirements) explicitly.
- Provide agent-usable CLI examples (
orcurl
) when requested.npx
4) Validate consistency
- Ensure message types and fields align with the envelope in
.references/protocol.md - Ensure endpoints link back to the discovery flow and offer/execute lifecycle.
5) Keep outputs agent-native
- Prefer declarative structures (intent, capabilities, proofs) over prose.
- Include provenance and evidence fields whenever output represents an execution.