parallel-coding-agent-loop

Ground implementation decisions in recorded ClimbHill resources and cited prior work.

Recombine proven parts of multiple evaluated attempts into a coherent follow-up attempt.

Review policy, evaluations, cost, lineage, and risk before recording a human promotion or rejection decision.

Prepare a repository for controlled recursive optimization by defining goals, agent guidance, policies, evaluations, resources, and protected surfaces.

Classify proposed paths and enforce denied or approval-required ClimbHill policy before edits.

Sample successful and failed historical ClimbHill Attempts before planning new work.

Analyze ClimbHill runs, attempts, lineage, costs, failures, and recurring constraints.

Convert evidence-backed ClimbHill experiment findings into reviewable GitHub issue proposals.

Produce evidence-backed run and attempt reports with evaluations, costs, risks, and promotion recommendations.

Run Perplexity Sonar and Sonar Deep Research for web-grounded answers with citations and bounded search context.

Build and validate custom-corpus inputs for deep research from Google Docs, YouTube, local files, S3, arXiv, databases, Slack, email, tickets, and similar…

Run Exa-backed deep research, neural search, and page-content retrieval with bounded result and page-read counts.