Dev orchestrates 6 AI skills for a test-fix pipeline, spending just $3/week on DeepSeek
Own-Awareness8037 · reddit · 2026-10-05
A Reddit user shares a hands-on workflow built with Claude Code's Skills mechanism: five skills for testing, bug fixing, data sanitization, deduplication, and test generation, plus an orchestrator skill, each with its own references file and scripts to keep runs deterministic and token-efficient.
He built them with DeepSeek 4 Pro and ran them on 4.1 Flash, consuming 250k tokens for about $3 a week with decent results. But when he orchestrated the identical setup via Antigravity-CLI (Gemini-based), Gemini 3.1 Pro produced poor results under the same prompts and token budget. He's asking for advice on how skills actually work, managing them in a Java Maven project, and whether paying for Claude is worth it.
More from coding & agent
- Agent fails: Dots subagent refuses user approval passed via main agent — xeophon · 2026-10-05
- Open-source App Store Connect CLI hits 7.7k stars, scriptable automation for iOS releases — rudrank · 2026-10-05
- Personal agents hackathon: 100+ ideas signal an agent explosion as infra matures — jamdac · 2026-10-05
- Browser agents' worst failure mode isn't hallucination, it's fake success — Comfortable_Oven6576 · 2026-10-05
- Power user review: ChatGPT Dots hit abuse-prevention limits even on the $200 plan — invertednz · 2026-10-05
- dots developer books daily user interviews to steer its personal assistant roadmap — pvncher · 2026-10-05