v0.1.0

What's Changed

• refactor: rename project from synth-bench to dataforge by @echobt in #1
• feat: Benchmark evaluation framework with agent-based task testing by @echobt in #2
• feat(benchmark): recalibrate timeouts and step requirements for longer task durations by @echobt in #3
• feat(openrouter): add reasoning model support for Kimi K2.5 by @echobt in #4
• fix(llm): improve JSON extraction for reasoning models by @echobt in #5
• feat(workspace): add advanced synthetic workspace generator with multi-agent debate system by @echobt in #6
• chore(build): remove stale Rust build artifacts from generated-datasets by @echobt in #7
• feat(dataset): add validated SWE-bench task dataset with 9 tasks across 3 difficulty levels by @echobt in #9
• chore: cluster indexation - AGENTS.md, git hooks, CI workflow, versioning by @echobt in #8
• feat(swe): harden pipeline quality gates and improve harness reliability by @echobt in #10
• fix(pipeline): validate test file references in exported tasks by @alpha1122x in #11
• feat(swe): fix harness for Docker agent execution and add test results by @alpha1122x in #12
• feat(swe): improve dataset generation pipeline with validation and progress monitoring by @echobt in #13
• feat(benchmark): add benchmark command with pipeline metrics and PR analysis by @echobt in #14
• feat(swe): improve pipeline concurrency, filtering, and pre-export validation by @echobt in #15
• feat(docker): 32GB RAM containers, pipeline concurrency config, containerized extraction, and tests by @echobt in #16
• fix(llm+swe): add tool_calls support to LiteLlmClient and fix pipeline bugs by @echobt in #17
• fix(swe): resolve tool server startup failures with port allocation, retries, and shell fallback by @echobt in #18
• fix(export): HF upload tasks/ dir, org repo creation, enrich concurrency by @alpha1122x in #19

New Contributors

• @echobt made their first contribution in #1
• @alpha1122x made their first contribution in #11

Full Changelog: https://github.com/CortexLM/swe-forge/commits/v0.1.0

Latest Release

This is an automatically updated release tracking the latest tagged version (v0.1.0).

Install with:

curl -fsSL https://raw.githubusercontent.com/CortexLM/swe-forge/main/install.sh | sh