ai-agent-guidelines

To enforce TDD for an autonomous AI agent, apply guidelines that mandate a Red-Green-Refactor cycle: write a failing test, implement minimal code to pass, refactor safely, then run all tests and report completion when green.

An autonomous AI agent can immediately execute bug fixes, refactoring updates, test additions, and documentation updates without confirmation, but requires approval for new files, schema changes, and deploy settings to ensure safe change scopes.

Context management for AI agents in software development involves defining specific context handling rules, enforcing completion reporting formats, and sending clear completion signals to ensure traceability and handoff readiness after automated edits.

Manual confirmation is required for AI agent repository edits when changes involve new file creation, schema modifications, or deploy settings, while dependency changes and quality assurance workflows need verification before commits.

Yes, the framework supports quality assurance and bug fix automation by applying a rules-driven decision framework that enforces test-driven workflows, distinguishes safe execution scopes, and requires test verification before commits.

The best way to structure completion reporting for autonomous AI agents is to enforce standardized reporting formats and clear completion signals, ensuring every automated task maintains traceability and review readiness.