{"slug": "belay-see-what-keeps-going-wrong-in-your-claude-code-and-codex-sessions", "title": "Belay – See what keeps going wrong in your Claude Code and Codex sessions", "summary": "Doplex Labs launched Belay, a local tool that scans existing Claude Code and Codex session histories to flag loops, stalls, and failed commands, then proposes reviewed rules and verification commands for the next run. In a preprint evaluation of approved lessons delivered to agents, tasks were completed in 11 of 16 cases with Belay versus 7 of 16 with a human-written instruction and 6 of 16 with no additional guidance. Belay stores session history, evidence, and guidance in an encrypted local database, requires no account, and does not upload data to Doplex Labs.", "body_md": "Across the sessions, the agent said done while the typecheck still failed. Belay groups the repeated pattern and keeps the proof attached.\n\nFor the next run, your agent receives a proposed rule and verification command. You review it before it has any authority.\n\nBuilt for the agents you already use.\n\nClaude Code\nCodex\n\nInstall once. Start with the sessions you already have.\n\n1\n\nYour history, in one place.\n\nBelay finds the coding sessions saved on your machine and loads them into a local timeline. You don't need to start a new run to find something useful.\n\nLoading your session history\n\nClaude Code\n\nCodex\n\n~/.belay/timeline.html\n\nYour sessions. On your machine.\n\n✓ Existing history loaded. New sessions added as you work.\n\n2\n\nFind what kept going wrong.\n\nBelay flags loops, stalls, and failed commands across your history. Open a session to see the work that went nowhere, and where you had to step in.\n\nBilling webhooksSession timeline\n\n14:02\n\nEdited six files across the APIpackages/api/src/billing/*\n\n14:31\n\nRan the unit tests. All passed.pnpm test billing → 48 passed\n\n14:33\n\nMarked the task complete. Never ran the typecheck.pnpm typecheck → 14 errors, left behind\n\n15:10\n\nYou stepped in.\"the build is red again\"\n\nThird session this week to end this way. Now your agent can see that.\n\n3\n\nSee what it cost you. And what to say instead.\n\nYour own agent debriefs each session: where your prompting or scoping cost time and money, and the exact message to send next time.\n\nBilling webhooks · DebriefWritten by your Claude Code\n\nYou named the files on your fourth message. Before that, the agent searched the whole repo three times.\n\n13:38–14:02 searching for the files14:02 edits and tests\n\nScopingabout 24 min · roughly $6\n\nWhat you did\nAsked to “fix the webhook retry bug” with no file names. The agent grepped three times and read 40 files before you pasted the paths.\n\nWhat an expert would have done\nNamed the two files and the failing test in the first message. You knew where the bug lived.\n\nSay this insteadThe retry bug is in packages/api/src/billing/webhooks.ts and retry.ts. The failing test is webhooks.test.ts. Fix it there, no need to search the rest of the repo.\n\n1 of 2 things to change · 1 to keepStays on this machine\n\n4\n\nHave your agent fix the cause.\n\nAsk your agent to look at the failed sessions. It reads Belay's evidence through MCP and proposes a real fix: a hook, a rule, a project instruction. You review it before anything changes.\n\nYour coding agentConnected to Belay\n\nWhy do my sessions keep ending with a broken build?\n\nThree times this week I marked work done without running the typecheck. Nothing in this project stops me. I can add a hook that runs it before I'm allowed to finish.\n\nBelay prepares reviewed guidance with the evidence, scope, and verification command attached. Your next agent starts with the lesson that matters instead of rediscovering it from scratch.\n\nMission Pack · typecheck before done\n\n1\n\nrelevant lesson ready for the next session\n\nOriginal evidence attachedReady\n\nScope and limits includedReady\n\nVerification command includedReady\n\nPreprint results\n\nLessons from past work helped on new tasks.\n\nTasks completed when a prior lesson was useful, across Claude Code and Codex.\n\n11/ 16Belay\n\n7/ 16Human-written instruction\n\n6/ 16No additional guidance\n\nThis evaluation measured approved lessons delivered to agents; session debriefs have not been evaluated yet. Broader evaluation is underway; full methodology and agent-level results will follow in a research preprint.\n\nBelay stores session history, evidence, and guidance in an encrypted local database and does not upload it to Doplex Labs. No account required. Read the trust details.\n\n↳\n\nLocal collection. Local storage.\n\nBelay reads the session files already on your machine and keeps its records there. It works offline.\n\n↳\n\nYour agent stays your agent.\n\nBelay has no model. Optional analysis uses your configured Claude Code or Codex provider, under the account and policies you already use.\n\n↳\n\nYour work keeps moving.\n\nBelay never blocks an agent action. If Belay stops, your agent keeps going.\n\nBelay for Teams · Enterprise\n\nComing soon.\n\nOne Belay across every machine at your company.\n\nFleet-wide patterns\n\nOne fix for everyone\n\nLocal-first, like solo\n\nJoin the waitlist\n\nCompany\n\nMake the next run a little wiser.\n\nLess going in circles. More getting somewhere. For you, and for your agent.", "url": "https://wpnews.pro/news/belay-see-what-keeps-going-wrong-in-your-claude-code-and-codex-sessions", "canonical_source": "https://getbelay.vercel.app/", "published_at": "2026-09-22 04:43:40+00:00", "updated_at": "2026-09-22 04:54:07.547416+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "developer-tools", "agent-protocols"], "entities": ["Belay", "Doplex Labs", "Claude Code", "Codex", "MCP"], "alternates": {"html": "https://wpnews.pro/news/belay-see-what-keeps-going-wrong-in-your-claude-code-and-codex-sessions", "markdown": "https://wpnews.pro/news/belay-see-what-keeps-going-wrong-in-your-claude-code-and-codex-sessions.md", "text": "https://wpnews.pro/news/belay-see-what-keeps-going-wrong-in-your-claude-code-and-codex-sessions.txt", "jsonld": "https://wpnews.pro/news/belay-see-what-keeps-going-wrong-in-your-claude-code-and-codex-sessions.jsonld"}}