AgentWeave v0.7.1

What's Changed

• docs: clarify PyPI installation and v0.7.0 release by @sauravsingla in #54
• chore: non-runtime hardening and release clarity by @sauravsingla in #55
• hardening: close pre-1.0 runtime, security, packaging and quality gaps by @sauravsingla in #56
• Improve newcomer onboarding and simplify the discovery header by @sauravsingla in #61
• docs: add provider-neutral local runtime example by @fatihcvs in #64
• Close file-backed SQLite connections deterministically by @fatihcvs in #62
• Add runnable example of duplicate tool names across MCP servers by @fatihcvs in #60
• docs: refresh contributor issue tiers by @sauravsingla in #69
• test: check repository-relative Markdown links by @fatihcvs in #75
• test: verify the plugin example contract offline by @fatihcvs in #74
• test: cover the offline BYOM example by @fatihcvs in #73
• docs: add examples index by @Zerro15 in #63
• Harden runtime replay safety, provenance, serialization, and CI by @sauravsingla in #77
• Release AgentWeave v0.7.1 by @sauravsingla in #78

New Contributors

• @fatihcvs made their first contribution in #64
• @Zerro15 made their first contribution in #63

Full Changelog: v0.7.0...v0.7.1

AgentWeave v0.7.0

What's Changed

• Add stage-specific failure analysis for Issue #29 by @sauravsingla in #30
• Add recovery and state-change stress tests for Issue #29 by @sauravsingla in #31
• Add fail-closed authorization and stage-drop observability for Issue #28 by @sauravsingla in #32
• Add adversarial exposure and authorization benchmark for Issue #28 by @sauravsingla in #33
• Add policy-first static scope filtering before routing for Issue #27 by @sauravsingla in #34
• Add portable tool-routing provenance for Issue #22 by @sauravsingla in #36
• Evaluate deferred tool search and progressive disclosure for Issue #35 by @sauravsingla in #37
• Harden trust boundaries and add adaptive hybrid routing for Issue #38 by @sauravsingla in #49
• Prevent credential forwarding across A2A redirect origins by @sauravsingla in #50
• Build plug-and-play secure AgentWeave runtime by @sauravsingla in #51
• Harden 0.7 runtime identity, authorization and integration lifecycle by @sauravsingla in #52
• Complete 0.7 release proof, scale, MCP E2E and live-provider protocol by @sauravsingla in #53

Full Changelog: v0.6.0...v0.7.0

AgentWeave v0.6.0 — Research Preview: Pre-Inference Routing for Tool-Rich Agents

AgentWeave v0.6.0 — Research Preview

AgentWeave is an open-source framework for deterministic pre-inference routing, policy-aware scope reduction, provenance, and reliability in tool-rich and multi-agent systems.

This release consolidates the current AgentWeave research and engineering foundation, including practical framework integrations, reproducible evaluation infrastructure, and a growing focus on failure analysis and recovery.

Highlights

• Pre-inference routing — reduces the tool or agent set exposed to the model before inference.
• Deterministic routing — supports reproducible tool and agent selection.
• MCP integration — demonstrates routing over MCP-style tool descriptors.
• LangGraph integration — applies AgentWeave routing before downstream graph execution.
• AutoGen integration — supports pre-selection of agents and actions before orchestration.
• Bring Your Own Model (BYOM) — enables alternative model integrations while preserving separation from the frozen research core.
• Policy-aware scope reduction — provides a foundation for controlling which capabilities become model-visible.
• Provenance and observability — supports analysis of routing decisions and downstream execution behavior.

Validation

The v0.6.0 release pipeline includes automated testing and package validation.

Current validation includes:

• 83 automated tests passing
• successful Python wheel build
• successful source distribution build
• successful twine check
• release-tag and package-version consistency checks

Research evaluation

AgentWeave includes a frozen BFCL-derived routing-pressure study investigating whether reducing the model-visible tool space can affect downstream function-calling reliability.

These experiments are research evaluations and are not official BFCL leaderboard results.

Historical benchmark evidence remains frozen and is not retroactively modified by newer framework integrations or experimental extensions.

Failure analysis and recovery

Recent external research feedback has motivated a deeper evaluation phase focused on:

• routing-stage failures;
• function-selection failures;
• multi-function composition failures;
• execution correctness;
• recovery after tool failure;
• state changes during execution;
• re-routing when previously pruned capabilities become necessary.

This work is tracked in Issue #29.

Evidence boundary

This release does not modify previously frozen benchmark results.

New integrations, BYOM functionality, recovery experiments, and future evaluation extensions remain separated from the historical experimental evidence to preserve reproducibility.

Status

Research Preview / Pre-release

AgentWeave is under active development. APIs, evaluation methodology, recovery behavior, and framework integrations may evolve before a production-stable release.

Feedback, reproducibility reports, issues, and research collaboration are welcome.