Draft protocol + Python reference implementation

Identity-aware delegation for multi-agent systems

A2A tells agents how to talk. LDP tells them which agent to call and whether to trust the answer. Expose delegate capability, cost, latency, provenance, and trust metadata so routers can make better decisions than skill-name matching alone.

Application — DCI (deliberative reasoning) optional Delegation Intelligence — LDP identity · routing · provenance · trust Agent Communication — A2A Tool Integration — MCP LDP here
37%
Token reduction with
semantic frames
Empirical, p=0.031 · paper
12x
Lower latency on easy tasks
via delegate specialization
Empirical, 3-delegate pool · paper
55%
Cost savings with
identity-aware routing
Simulated routing demo · code
0.91
Confidence tracking with
verified provenance
Protocol feature · per-response metadata

Smart routing vs blind routing

Three tasks, three delegates. Blind routing sends everything to the most expensive model. LDP matches task complexity to delegate capability.

Discovered 3 delegates:

  Fast Agent       gemini-2.0-flash       quality=0.60  cost=$0.001  p50=200ms
  Balanced Agent   claude-sonnet-4-6      quality=0.82  cost=$0.008  p50=1200ms
  Deep Agent       claude-opus-4-6        quality=0.95  cost=$0.025  p50=3500ms

── Blind Routing (skill-name only) ──────────────────────────

  Task (easy  )  ->  Deep Agent       cost=$0.025  latency=3496ms  <- overkill
  Task (medium)  ->  Deep Agent       cost=$0.025  latency=3493ms  <- expensive
  Task (hard  )  ->  Deep Agent       cost=$0.025  latency=3755ms  <- correct

  Total: $0.075  |  10,744ms

── LDP Routing (identity-aware) ─────────────────────────────

  Task (easy  )  ->  Fast Agent       cost=$0.001  latency=200ms   <- right-sized
  Task (medium)  ->  Balanced Agent   cost=$0.008  latency=1108ms  <- right-sized
  Task (hard  )  ->  Deep Agent       cost=$0.025  latency=3582ms  <- right-sized

  Total: $0.034  |  4,890ms

── Comparison ──────────────────────────────────────────────

  Cost savings:    55% ($0.075 -> $0.034)
  Latency savings: 54% (10,744ms -> 4,890ms)

── Provenance (LDP exclusive) ───────────────────────────────

  produced_by:  ldp:delegate:deep-01
  model:        claude-opus-4-6
  confidence:   0.91
  verified:     true
  payload_mode: semantic_frame (37% fewer tokens than text)

Delegation intelligence for agent protocols

LDP complements, not replaces, existing protocols. Use A2A for message exchange, MCP for tools, and LDP for delegate identity, routing, provenance, and trust.

Rich delegate identity

Model family, quality scores, reasoning profiles, cost and latency hints. The information routers actually need — not just a name and skill list.

Progressive payload modes

Six encoding modes from text to semantic frames. Negotiate the most efficient format automatically. 37% token reduction measured empirically.

Structured provenance

Every response carries who produced it, which model, confidence score, and verification status. The accountability infrastructure multi-agent systems need.

Trust domains

Protocol-level security boundaries beyond transport auth. Cross-domain delegation requires explicit trust. Application-level enforcement, not just TLS.

Intelligent routing

Route by quality, cost, latency, or balanced score. Send easy tasks to fast, cheap models. Hard tasks to capable, expensive ones. Right-size every delegation.

Governed sessions

Persistent context across multi-round delegations. No re-transmitting conversation history. Eliminates quadratic token overhead in long interactions.

Build a delegate in 10 lines

Install
pip install ldp-protocol
Create a delegate
from ldp_protocol import LdpDelegate, LdpCapability, QualityMetrics

class MyDelegate(LdpDelegate):
    async def handle_task(self, skill, input_data, task_id):
        return {"answer": "42"}, 0.95  # (output, confidence)

delegate = MyDelegate(
    delegate_id="ldp:delegate:my-agent",
    name="My Agent",
    model_family="claude",
    model_version="claude-sonnet-4-6",
    capabilities=[LdpCapability(
        name="reasoning",
        quality=QualityMetrics(quality_score=0.85),
    )],
)
delegate.run(port=8090)
Route intelligently
from ldp_protocol import LdpRouter, RoutingStrategy

async with LdpRouter() as router:
    await router.discover_delegates([
        "http://fast-model:8091",
        "http://deep-model:8092",
    ])
    result = await router.route_and_submit(
        skill="reasoning",
        input_data={"prompt": "Analyze..."},
        strategy=RoutingStrategy.QUALITY,
    )

Grounded in empirical evaluation

LDP is research-backed, not just an API spec. Documented in public research papers with reproducible experiments and open-source evaluation code.

Notable finding

Accurate provenance did not significantly improve synthesis quality over no provenance at all. But noisy provenance — unverified self-reported confidence — actively harmed quality, doubling output variance. This is why LDP includes explicit verification status, not just confidence scores.

Read the full analysis
LDP: An Identity-Aware Communication Protocol for Multi-Agent LLM Systems
Sunil Prakash · arXiv:2603.08852 · 2026
arXiv · PDF · Experiment code
DCI: Deliberative Collective Intelligence
Sunil Prakash · arXiv:2603.11781 · 2026
arXiv · Code
DCI provides the reasoning layer; LDP provides the delegation protocol.

Start building with LDP

Route to the right agent. Know what produced the answer.