{"slug": "show-hn-a-simple-decision-layer-for-coding-agents", "title": "Show HN: A simple decision layer for coding agents", "summary": "Developer smarchone released Signals, a Claude Code skill installed via `npx skills add smarchone/signals -g`, that records the decisions made during coding sessions and carries them across sessions so agents stop re-litigating settled calls. The tool stores each decision with its rationale and what was ruled out, and surfaces only new decisions that are hard to undo or felt by users; commands include `/signals why sqlite?`, `/signals build the first slice` and `/signals exec drop sqlite for pg`. Setup takes roughly one minute and requires Claude Code.", "body_md": "# [~/ smarchone/signals](https://github.com/smarchone/signals)\n\ntl;drA Claude Code skill. It captures the decisions you make as you talk and build, across sessions, and asks you only about new ones that are hard to undo or felt by users.\n\nWhat are you steering?\n\n[the thesis →](blog.html)\n- **Sessions** → none of them knows what the others decided.\n- **Prompts** → prompt hell. You re-steer the same calls, prompt after prompt.\n- **Specs** → spec-driven dev. The spec is stale by build two. Back to prompting.\n- **Decisions** → signals. Decided once, it steers every build after.\n\nBig calls only\n\n## Decisions, not diffs.\n\nOnly the ones hard to undo, or felt by users.\n\nDecided stays decided\n\n## Each keeps its why.\n\nAnd what was ruled out, so agents stop re‑arguing it.\n\nExec level\n\n## You stay at the top.\n\nNever variable names. Single player, for now.\n\nyou talk\n\n*→*agent builds*→*you approve new decisions*→*remembered\nGet started · **~1 min** · needs Claude Code\n\n``` bash\n$ npx skills add smarchone/signals -g   # then, in Claude Code:\n> /signals why sqlite?                  # answers from your decisions\n> /signals build the first slice        # builds what you discussed\n> /signals exec drop sqlite for pg      # the panel argues, you decide\n```\n\nmade with ♥ · [@smarchone/signals](https://github.com/smarchone/signals)", "url": "https://wpnews.pro/news/show-hn-a-simple-decision-layer-for-coding-agents", "canonical_source": "https://smarchone.github.io/signals/", "published_at": "2026-11-22 19:14:59+00:00", "updated_at": "2026-11-22 19:42:33.311702+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "developer-tools"], "entities": ["smarchone", "Signals", "Claude Code", "GitHub"], "also_reported_by": [], "takeaway": "Signals is a Claude Code skill, installed with `npx skills add smarchone/signals -g`, that persists approved architectural decisions and their rejected alternatives across sessions so coding agents stop re-arguing settled choices.", "recommendation": null, "alternates": {"html": "https://wpnews.pro/news/show-hn-a-simple-decision-layer-for-coding-agents", "markdown": "https://wpnews.pro/news/show-hn-a-simple-decision-layer-for-coding-agents.md", "text": "https://wpnews.pro/news/show-hn-a-simple-decision-layer-for-coding-agents.txt", "jsonld": "https://wpnews.pro/news/show-hn-a-simple-decision-layer-for-coding-agents.jsonld"}}