Skip to content

CHANGELOG.md

Latest commit

 

History

History
83 lines (61 loc) · 3.51 KB

CHANGELOG.md

File metadata and controls

83 lines (61 loc) · 3.51 KB
Raw

Changelog

All notable changes to the Capability Discovery Protocol (CDP) will be documented in this file.

The format follows Keep a Changelog, and this project adheres to Semantic Versioning.

Naming note: This project was originally released as ADP (Agent Discovery Protocol). It was later renamed CFP (Capability Fabric Protocol) during capability-centric evolution, and is now called CDP (Capability Discovery Protocol).

0.1.1 — 2026-03-30

Changed

  • Finalized rename from ADP/CFP to CDP (Capability Discovery Protocol)
  • Standardized naming across repository (README, docs, schemas, examples)
  • Updated project positioning to capability-centric discovery
  • Added and refined architecture diagrams (indexed, federated, hybrid, agent-less flows)
  • Expanded documentation for agent-less execution mode
  • Clarified Provider → Skill → Offering model across docs and examples

Fixed

  • Corrected Makefile and tooling references after package rename (adp → cdp)
  • Updated lint, format, and validation commands to use cdp/
  • Ensured CI workflow uses correct paths
  • Removed stale references to adp/ in code and documentation

Documentation

  • Updated README to reflect CDP naming and positioning
  • Added docs/no-agent-mode.md
  • Renamed docs/adp-context.md → docs/cdp-context.md
  • Improved protocol overview and architecture sections
  • Added clear explanation of CDP vs ADP vs CFP evolution

Notes

  • This is a stabilization and consistency release after the CDP transition
  • No functional breaking changes to discovery behavior
  • Python package now uses cdp instead of adp

0.1.0 — 2024-01-15

Initial public release of the Capability Discovery Protocol reference implementation.

Added

Protocol

  • DiscoveryIntent schema with 6 categories: product, service, agent, data, compute, api
  • ProviderDescriptor and OfferingDescriptor schemas
  • 14-stage discovery pipeline: parse → normalize → extract constraints → retrieve candidates → deterministic filter → semantic matching → capability validation → policy filtering → ranking → explanation generation → federation merge → deduplication → audit logging → response generation
  • 4 ranking profiles: default, cost_optimized, latency_optimized, trust_optimized
  • Federation support for distributed CDP node queries
  • Policy engine with trust-level enforcement

Reference Implementation (Python 3.11+)

  • FastAPI server with /discover, /register/provider, /register/offering, /health endpoints
  • In-memory registry with TTL-based expiry
  • Keyword-based semantic matching
  • Weighted scoring with per-field explainability
  • Audit logging via DiscoveryAuditRecord

Schema & Vectors

  • 9 JSON Schema Draft 7 definitions in schema/
  • 10 canonical protocol vectors in protocol-vectors/

Tests

  • Unit, functional, schema validation, protocol vector, and interoperability test suites

Documentation

  • Architecture overview, wire spec, discovery model, federation model, ranking model, security model
  • Quickstart guide and example walkthroughs

Known Limitations

  • Semantic matching is keyword-based (no vector embeddings in v0.1)
  • Registry is in-memory only (no persistence layer)
  • No built-in authentication (must be added at deployment)