Manual QA harness for the senpi coding agent itself. MUST USE after changing packages/ai, packages/agent, packages/coding-agent, or packages/tui — a green typecheck and `npm test` are NOT QA. Drives the real CLI from source in an isolated sandbox (never touches ~/.senpi or real credentials) across four channels: remote RPC (--mode rpc JSONL stdio), TUI smoke (node-pty on Windows, tmux on POSIX), mock loop (a local fake model server for deterministic, zero-token agent-loop runs), and CLI smoke (--help/--print/--list-models). Every helper ships a --self-test. Use whenever someone says qa senpi, test the agent, verify my change, rpc qa, tui qa, mock-loop qa, smoke the cli, or needs evidence that an agent-loop, tool, keybinding, or provider change works end to end. Capture evidence to local-ignore/qa-evidence/.
Read before a js eval cell that spawns a process, server or PTY, installs a package, or starts a long run, and before any bun -e, script, server, CLI, test, typecheck, format, bundle, or package-management work: this session's eval js kernel runs Bun 1.4+. Bun 1.4.x builtins replace 20+ npm deps and dev tools — check here BEFORE installing or running tsc/tsgo, prettier/oxfmt, sharp, puppeteer/playwright (scraping), cheerio, marked, node-cron, node-pty, execa/zx, concurrently, serve-static, tar, json5, fast-xml-parser, string-width, dotenv, keytar, aws-sdk (S3/R2) — and before writing Bun.$ / Bun.spawn / bun test code: the traps that cost past sessions hours are listed here. Triggers: eval js, bun, Bun.$, Bun.spawn, Bun.serve, bun test, bun check, bun fmt, bun build, bun install, image resize, headless browser, markdown, cron, PTY, secrets.
MUST read before generating images. Prompt-crafting guide for gpt-image-2.6 covering tool routing (native image_generation server tool vs the generate_image tool), model and quality selection, prompt structure, exact text rendering, reference-image editing, transparent assets, output formats, and multi-turn refinement.
