Topic archive
How to structure a coding agent CI repair loop: detect a failing CI run, hand the failure context to an agent, validate the fix, and gate the PR without losing evidence or spinning into runaway retry cycles.
Practical cost-control patterns for teams running coding agents through a model gateway, covering spend guards, per-agent token budgets, CI-integrated spend checks, and verification steps to confirm your gateway enforces the limits you configure.
A practical guide to choosing and configuring model routes for coding agent workflows, covering task-type mapping, instruction-file conventions, CI integration, and gateway setup.
Coding agents read instruction files — AGENTS.md, CLAUDE.md, .github/copilot-instructions.md, and memory files — to understand what tools, workflows, and conventions are available. When those files drift from the actual runtime environment, agents make confidently wrong decisions. This guide explains what prompt registry drift is, how it appears across popular agent runtimes, and how to detect and prevent it with CI-backed drift checks.
A source-checked migration checklist for moving coding agents toward an OpenAI-compatible gateway while preserving repository instructions, secret boundaries, endpoint-family tests, and smoke-test evidence.
A practical GitHub Actions triage workflow for asking a coding agent to reproduce, explain, patch, and hand off CI failures.
A source-backed operating guide for routing coding agent writer, reviewer, critic, and fallback calls through a model gateway without confusing Responses, Chat Completions, Claude messages, secrets, or publish gates.
A practical rule set for assigning parallel coding agents to separate Git worktrees, branches, and pull request handoffs.
A practical workflow for checking coding agent outputs against repository instructions, official documentation, diffs, and a short evidence log.
A practical checklist for handing a coding-agent pull request to reviewers with scope, checks, instruction context, risk notes, and clean pass/fail records.
A practical boundary plan for giving coding agents useful repository context while keeping credentials, deployment rights, and CI tokens under human control.
A practical pattern for writing task briefs that give coding agents enough repository rules, acceptance criteria, and review handoff structure to produce changes a human can inspect.
A practical guide to writing repository-level instructions, agent rules, and task briefs that help coding agents work safely in real projects.