Spec Kit Agents: Context-Grounded Agentic Workflows
This addresses context blindness in spec-driven development for AI coding agents, offering incremental improvements in repository compatibility and benchmark performance.
The paper tackled the problem of AI coding agents hallucinating APIs and violating architecture in large repositories by introducing Spec Kit Agents, a multi-agent pipeline with context-grounding hooks, which improved judged quality by +0.15 on a 1-5 score and achieved 58.2% Pass@1 on SWE-bench Lite.
Spec-driven development (SDD) with AI coding agents provides a structured workflow, but agents often remain "context blind" in large, evolving repositories, leading to hallucinated APIs and architectural violations. We present Spec Kit Agents, a multi-agent SDD pipeline (with PM and developer roles) that adds phase-level, context-grounding hooks. Read-only probing hooks ground each stage (Specify, Plan, Tasks, Implement) in repository evidence, while validation hooks check intermediate artifacts against the environment. We evaluate 128 runs covering 32 features across five repositories. Context-grounding hooks improve judged quality by +0.15 on a 1-5 composite LLM-as-judge score (+3.0 percent of the full score; Wilcoxon signed-rank, p < 0.05) while maintaining 99.7-100 percent repository-level test compatibility. We further evaluate the framework on SWE-bench Lite, where augmentation hooks improve baseline by 1.7 percent, achieving 58.2 percent Pass@1.