The Coding Agent That Can't Grade Its Own Homework

Description and use

The agent‑loop framework splits code generation into stages with file handoffs and independent verification, eliminating self‑confirmation and providing auditability.

Problem

Coding agents often self‑grade, leading to false “All tests pass” messages and noisy, biased outputs.

Outcome

Agent‑loop delivers deterministic, verifiable code changes with a clear audit trail, preventing self‑confirmation bias and ensuring tests are actually run.

Workflow

• 1. Triage: classify task into S, M, or L tier
• 2. Analyze (M/L): generate analysis.md
• 3. Plan (M/L): produce plan.md with runnable check
• 4. Implement (S/M/L): create implementation.md
• 5. Verify: run independent verifier on plan and diff, output PASS/FAIL
• 6. Debug (on FAIL): up to 3 iterations, each with fresh context and re‑verification

Evidence

The verifier never sees the reasoning

Source

The debug loop caps at 3 iterations

Source