Skip to content

Releases: limecloud/agentruntime

Agent Runtime v0.3.9

08 May 09:29

Choose a tag to compare

Agent Runtime v0.3.9

Agent Runtime v0.3.9 is a patch release that refreshes the Agent standards ecosystem after Agent Context became a current published standard.

Highlights

  • Adds Agent Context to the current standards table and boundary map.
  • Removes Agent Context from the future-candidate list.
  • Adds Agent Context to README related-standard links.
  • Updates llms.txt, llms-full.txt, and public LLM copies so AI clients can discover Agent Context.
  • Adds version snapshot pages for v0.3.9.

Links

Agent Runtime v0.3.8

08 May 08:55

Choose a tag to compare

Agent Runtime v0.3.8

Agent Runtime v0.3.8 is a patch release that refreshes the Agent standards ecosystem after Agent Tool became a current published standard.

Highlights

  • Adds Agent Tool to the current standards table and boundary map.
  • Removes Agent Tool from the future-candidate list.
  • Adds Agent Tool to README related-standard links.
  • Updates llms.txt, llms-full.txt, and public LLM copies so AI clients can discover Agent Tool.
  • Adds version snapshot pages for v0.3.8.

Links

Agent Runtime v0.3.7

08 May 08:25

Choose a tag to compare

Agent Runtime v0.3.7

Agent Runtime v0.3.7 is a patch release that refreshes the Agent standards ecosystem after Agent Artifact became a current published standard.

Highlights

  • Adds Agent Artifact to the current standards table and boundary map.
  • Removes Agent Artifact from the future-candidate list.
  • Adds Agent Artifact to README related-standard links.
  • Updates llms.txt, llms-full.txt, and public LLM copies so AI clients can discover Agent Artifact.
  • Adds version snapshot pages for v0.3.7.

Links

Agent Runtime v0.3.6

08 May 07:50

Choose a tag to compare

Agent Runtime v0.3.6

Agent Runtime v0.3.6 is a patch release that refreshes the Agent standards ecosystem after Agent Policy became a current published standard.

Highlights

  • Adds Agent Policy to the current standards table and boundary map.
  • Removes Agent Policy from the future-candidate list.
  • Adds Agent Policy to README related-standard links.
  • Updates llms.txt, llms-full.txt, and public LLM copies so AI clients can discover Agent Policy.
  • Adds version snapshot pages for v0.3.6.

Links

Agent Runtime v0.3.5

08 May 05:09

Choose a tag to compare

Agent Runtime v0.3.5

Agent Runtime v0.3.5 fixes repository-base homepage asset links. The localized home pages now keep their home layout while LLM entrypoint links resolve under the project site path and the navigation logo loads from the correct public asset path.

Highlights

  • Fixes localized homepage LLM entrypoint links for GitHub Pages repository-base deployments.
  • Fixes the documentation logo asset path under repository-base deployments.
  • Keeps the localized home page layout correction from v0.3.4.
  • Keeps the core Agent Runtime specification compatible with v0.3.4.
  • Updates version snapshots and package metadata.

Validation

  • VITEPRESS_BASE repository-base build
  • Localized homepage layout checks
  • LLM file consistency checks
  • git diff --check
  • npm audit --omit=dev
  • npm pack --dry-run

Agent Runtime v0.3.4

08 May 04:59

Choose a tag to compare

Agent Runtime v0.3.4

Agent Runtime v0.3.4 fixes the Simplified Chinese homepage and aligns both localized index pages with the home landing layout.

Highlights

  • Fixes Simplified Chinese homepage rendering by using proper VitePress home layout for localized index pages.
  • Refines English and Simplified Chinese home pages with concise hero actions, quick links, ecosystem links, and LLM entrypoints.
  • Fixes repository-base GitHub Pages logo loading.
  • Keeps the core Agent Runtime specification compatible with v0.3.3.
  • Updates version snapshots and package metadata.

Validation

  • npm run build
  • Repository-base VITEPRESS_BASE build
  • npm audit --omit=dev
  • npm pack --dry-run
  • LLM file consistency checks
  • Localized homepage layout checks

Agent Runtime v0.3.3

08 May 04:43

Choose a tag to compare

Agent Runtime v0.3.3

Agent Runtime v0.3.3 adds public ecosystem exposure across the Agent standards family. Each standard now links to the other published standards and documents future standard candidates.

Highlights

  • Adds an Agent standards ecosystem page under reference/agent-ecosystem.
  • Links Agent Knowledge, Agent UI, Agent Runtime, and Agent Evidence from the public docs.
  • Adds English and Simplified Chinese ecosystem pages.
  • Adds ecosystem links to llms.txt and llms-full.txt.
  • Updates version snapshots and package metadata.

Validation

  • npm run build
  • Repository-base VITEPRESS_BASE build
  • npm audit --omit=dev
  • npm pack --dry-run
  • LLM file consistency checks
  • Forbidden local-coupling scan

Agent Runtime v0.3.2

08 May 04:00

Choose a tag to compare

Agent Runtime v0.3.2

Agent Runtime v0.3.2 adds LLM-friendly documentation entrypoints. It publishes concise and full-context markdown files at the repository root and the documentation site root so AI clients can discover the runtime standard directly.

Highlights

  • Adds llms.txt as the concise LLM navigation index.
  • Adds llms-full.txt as a concatenated English core documentation file with source URLs.
  • Adds compatible llm.txt and llm-full.txt aliases.
  • Publishes the same files through docs/public/ so GitHub Pages serves them from the site root.
  • Includes the LLM entrypoint files in the package manifest.

Validation

  • npm run build
  • VITEPRESS_BASE="/agentruntime/" npm run build
  • npm audit --omit=dev
  • npm pack --dry-run
  • Root/public LLM file consistency checks
  • Forbidden coupling scan outside generated output

Agent Runtime v0.3.1

08 May 03:43

Choose a tag to compare

Agent Runtime v0.3.1

Agent Runtime v0.3.1 strengthens A2A alignment without turning Agent Runtime into an A2A implementation. A2A is now documented as the peer-agent interoperability reference for remote delegation, while Agent Runtime remains the owner of local execution facts, attempts, permissions, evidence, and recovery.

Highlights

  • Added official A2A references: latest protocol specification, Google launch post, Linux Foundation project announcement, and A2A GitHub project.
  • Added an A2A boundary map for Agent Card, A2A task/context ids, messages, artifacts, streaming, push notifications, and in-task authorization.
  • Expanded remote channels with native_protocol, peer-agent roles, native peer ids, remote task refs, and peer task lifecycle events.
  • Added A2A peer-task fields including remote_task_id, remote_context_id, agent_card_ref, delivery_mechanism, native_status, and remote_artifact_refs.
  • Clarified that A2A messages are communication or interaction facts, while durable outputs should be artifact refs linked to tasks.

Validation

  • npm run build
  • VITEPRESS_BASE="/agentruntime/" npm run build
  • npm audit --omit=dev
  • npm pack --dry-run
  • Forbidden coupling scan outside generated output

Agent Runtime v0.3.0

08 May 03:27

Choose a tag to compare

Agent Runtime v0.3.0

Agent Runtime v0.3.0 adds Agent Task as a first-class execution object. It turns tasks from loose UI cards, checklist items, or background labels into durable runtime truth with lifecycle, attempts, task graph relationships, progress, delivery state, and recovery semantics.

Highlights

  • Agent Task is now a first-class contract with objective, owner, constraints, acceptance, progress, attempts, and relationships.
  • task, run, step, job, and todo are clearly separated so real execution does not collapse into a single message or checklist.
  • Task status now covers waiting input, waiting permission, waiting resource, blocked, paused, retrying, cancelling, timed out, lost, completed, archived, stale, and unknown states.
  • Public event schema now includes task orchestration and attempt lifecycle events such as task.progress, task.delegated, task.dependency.updated, and task.attempt.*.
  • Public snapshot schema now includes task records, attempts, relationships, summaries, blocked tasks, graph refs, and delivery state.
  • Control plane, telemetry, state snapshots, quickstart, examples, acceptance scenarios, glossary, and source analysis now carry task ids and run ids.

Validation

  • npm run build
  • VITEPRESS_BASE="/agentruntime/" npm run build
  • npm audit --omit=dev
  • npm pack --dry-run
  • JSON Schema parse and task event/schema coverage checks
  • Forbidden coupling scan outside generated output