Headron Adds Change-Aware Verification for AI Agents

Headron targets the most dangerous weakness in autonomous coding: confident completion claims backed by incomplete evidence. Its approach is practical and valuable, though deterministic checks still cannot prove semantic correctness on their own.

• –Change-aware verification focuses checks on modified files while preserving a full final verdict.
• –`headron port` routes agents to the context and obligations relevant to each file, reducing unnecessary prompt and repository overhead.
• –Explicit `SKIP`, `NOT RUN`, and `REUSED` states prevent missing or cached evidence from masquerading as fresh success.
• –Commit hooks, CI, MCP, and multiple agent adapters make the contracts apply across tools rather than locking teams into one coding agent.
• –Project-published metrics are promising, but independent validation and stronger semantic checks will determine whether Headron can become production-grade infrastructure.