AI Agent Workflow Collection

Create concise role-specific Codex window prompts and current-window handoffs with an architecture-first gateway and role-window registry. Use when the user asks for an architecture/development/UI-PPT/UI-Frontend/video/ops/security/testing/QA/document-delivery/WeChat Official Account publishing/Xiaohongshu/personal knowledge-base window, says to inherit/reset/continue the current role, asks for the next-window prompt, wants a requirement routed through architecture before deciding whether to open other role windows, wants existing roles reused instead of recreated, wants testing test-case/report prompts, wants delivery-document prompts for requirements/contracts/acceptance/docs, wants content publishing roles, wants personal knowledge-base organization, or wants security-audit prompts that delegate to the appropriate security skill.

Read-only first workflow for diagnosing application incidents by collecting service, log, network, configuration, and functional evidence before proposing any fix.

Use for user-authorized black-box web security testing, especially when asked to act like a hacker, test a public site for data leakage or unauthorized modification, review exposed JS/API surfaces, login brute-force protections, CORS/security headers, or produce a penetration-style report. Triggers include 黑盒安全测试, 渗透测试, 扮演黑客, 漏洞测试, 数据泄露, 未授权访问, 登录爆破防护, 公网网站安全报告.

Use when the 文档/交付 role maintains client-facing project delivery document packages, especially Word/DOCX deliverables such as delivery lists, demo scripts, acceptance forms, change confirmations, quote notes, contract-adjacent materials, README/docs indexes, and internal handoff notes.

GStack design-html method for producing or planning quick HTML prototypes, design boards, or rendered visual explorations for UI review.

GStack document generation method for turning implementation, product, or operational work into durable docs with clear audience and ownership.

GStack guard method for strengthening scope boundaries, forbidden actions, stop conditions, and validation gates for an agent or workstream.

GStack health-check method for broad project or repository health review across tests, docs, DX, release readiness, and maintainability.

GStack developer-experience plan review for setup, local commands, docs, test loops, onboarding friction, and maintainability before implementation.

GStack plan-tune method for calibrating how many questions to ask, autonomy level, stop conditions, and decision style for a workstream or user preference.

GStack spec method for turning fuzzy intent into an executable spec with goals, non-goals, acceptance criteria, risks, and downstream role boundaries.

GStack unfreeze method for safely resuming a previously frozen task with explicit resume criteria, allowed changes, and verification gates.

Generate Guizang-style social card image sets and WeChat official account cover pairs from articles, scripts, screenshots, product notes, subtitles, or photos. Use when the user asks for 小红书图文, Rednote/Xiaohongshu images, social cards, carousel images, 3:4 covers, 微信公众号封面, WeChat 21:9 + 1:1 covers, Swiss Style, or magazine-style social images.

Diagnose Hermes cron jobs that ran successfully at scheduler level but produced empty, missing, or non-actionable user output.

Diagnose and safely document Hermes cron script interpreter mismatches, especially when shell scripts are executed as Python or Python wrappers are executed as shell.

Read-only workflow for discovering uploaded deployment artifacts, comparing them with current deployment state, and preparing a safe update and rollback plan.

Create cute 3D toy assets from photo-reference manifests. Use when Codex needs to turn scene photos into 2D toy concept prompts, choose a photo-to-3D route, produce or plan GLB assets for Three.js, or maintain a repeatable personal-gift/toy-character generation workflow.

Generic read-only post-deployment verification workflow for confirming service health, artifact activation, routing, logs, and rollback signals after a deployment action.

Generic read-only checklist for preparing an application deployment by verifying host, artifacts, backups, service state, configuration, dependencies, and rollback readiness before any write action.

Read-only diagnosis workflow for Python services whose external HTTP or WebSocket readiness differs between manual tests and service runtime behind a proxy.

Read-only first troubleshooting workflow for Python project deployments, including systemd startup failures, dependency issues, encoding problems, permissions, proxy setup, and readiness gates.

Create or update project test-case assets and test reports. Use when Codex is asked to generate test cases in Excel, produce a Word/DOCX test report, summarize automated test execution, turn repository tests into a formal testing artifact, or沉淀/复用 a project testing documentation workflow.

Operate the user's WeChat Official Account AI application content workflow. Use when working in an accounts/content repository on WeChat articles about AI applications, Agents, MCP, Skills, tool calling, workflows, open-source AI tooling updates, weekly AI application digests, image-rich WeChat article formatting, draft-box API updates, or local WeChat content handoff. Do not use for Xiaohongshu or Xianyu work.

Read and analyze Xiaohongshu/Rednote comments from the user's logged-in Chrome session. Use when the user asks to crawl, collect, export, summarize, classify, or use 小红书评论区 / XHS comments for content planning, follow-up posts, audience research, reply strategy, or 引流文案. Works best for already-open note pages or creator-manager note links.