common

Orchestrate production-ready LiveStacks solutions from PRDs, working PRDs, or minimal business-first inputs such as industry, pain point, category, Oracle feature mapping, deployment target, and implementation notes. Use when Codex needs to turn a business or industry pain point, PRD, workbook-style brief, or solution concept into a secure, portable, Oracle Database-first LiveStacks application bundle with architecture, data design, ORDS-mediated APIs, container artifacts, implementation guidance, documentation, and customer rebuild instructions.

Create clean distributable ZIP archives from files or project directories while excluding macOS metadata, Python caches, VCS metadata, editor temporary files, and other generated artifacts. Use when packaging a project, source tree, handoff bundle, deployment archive, or any ZIP that must not contain .DS_Store, __MACOSX, __pycache__, resource forks, or similar clutter.

Disable Polly retrieval and memory writes for the current local Git clone while preserving its canonical repository and quiet-mode settings. Use when a repository should no longer participate in Polly or was enabled unintentionally.

Enable Polly for the current clone or fork and configure its canonical upstream GitHub repository using local git config. Use when opting a repository into Polly or correcting fork-to-upstream memory resolution.

Diagnose Polly enrollment, repository enablement, secure token availability, canonical repository resolution, and server health. Use when Polly context is missing or setup needs verification.

Save a durable developer-only project memory in Polly without tying it to the current Codex session or sharing it with collaborators. Use when the user wants a personal decision, hypothesis, reminder, preference, or continuity note available in future sessions for the same canonical repository.

Pause, resume, or inspect Polly memory writes for the current git repository while keeping context retrieval active. Use before test prompts, experiments, sensitive exploratory work, or any session whose checkpoints and manual memories must not be stored.

Share a coding decision, constraint, assumption, blocker, or progress note with collaborators through Polly. Use when the user explicitly wants collaborators to receive a durable piece of project context.

Enroll this Codex device with an administrator-issued Polly code and store the permanent device token in the operating system credential store. Use when setting up Polly for the first time or replacing a revoked device token.

Create or update Oracle LiveLabs workshops from ideas or source materials such as blogs, product docs, web pages, Jupyter notebooks, PowerPoint decks, demos, or mixed source bundles. Use when you need a lean core skill that can start from source-only inputs, scaffold workshop structure, author markdown, optionally add FreeSQL runnable content, and validate against LiveLabs markdown rules.

Build, update, scaffold, and validate LiveStack demo guides that act as scene-by-scene runbooks for a working LiveStack application. Use when a LiveStack bundle needs a guide/ folder with desktop, sandbox, and tenancy workshop variants, real scene walkthrough labs, expected outcomes, screenshots, manifests, and portable download/run instructions aligned to the actual app.

Draft, restructure, validate, and iteratively refine Confluence storage-format pages. Use for new Confluence pages, versioned page workspaces, summary-first templates, storage markup, tabs, status legends, expand sections, duplicate-content QA gates, task reports, and targeted refinement questions.

Portable Oracle JET / Redwood app UI skill. Use for Oracle JET framework, Redwood theme, Oracle Sans typography, and JET glyph guidance using only packaged local references and assets.

Create and manage durable Codex task plans, PlanOps project folders, plan iterations, execution briefs, validation matrices, QA plans, status/resume files, decision logs, and targeted review questions for complex implementation, architecture, troubleshooting, documentation, cleanup, and research work.

Draft, revise, and render email newsletters in Markdown using emailmd. Use when a user asks to create a newsletter, email campaign, announcement email, internal update email, product/news roundup, or any reusable HTML email where the final deliverables should include both the Markdown source and the rendered HTML. Before rendering, verify the `emailmd` package is available; if not, install it with `npm install emailmd`.

Use when working with `freesql.com` / Oracle FreeSQL in a real browser: writing or running SQL, PL/SQL, Quick SQL, or JavaScript; exploring schemas and tutorials; capturing result screenshots; generating share links, 'Run in FreeSQL' buttons, or embedded editor iframe code; or troubleshooting FreeSQL UI behavior.

Launch, install, verify, or help operate the Outbound PM Operations Hub Jira dashboard. Use when the user asks to open the Jira dashboard, start the dashboard, fix live ticket fetching, share or install the dashboard package, or work with the Outbound PM operating hub.

Create quizzes and gamify Oracle LiveLabs workshops. Use when Codex needs to generate quiz questions from workshop content automatically, add a scored quiz as a new lab at the end of a workshop, append a quiz to a FastLab markdown, or distribute quiz blocks across existing lab markdown at valid section boundaries while preserving instructional content verbatim and updating the named manifest when required.

Convert any existing Oracle LiveLabs workshop into an industry-specific variant from one simple prompt. Use when Codex needs to inspect a source workshop, infer lab order and manifest structure, map it to a target industry, rewrite the labs, run strict LiveLabs and prose checks, catch leftover source vocabulary, verify launch files, and return a QA summary.

Package a local industry LiveStack application into a portable INPUT_ZIP for LiveLabs Workshop Author 26.5.2. Use when given a stack path and output directory and needing database-source, frontend-source, and backend-source or api-map.md without hard-coded local paths.

Team writing standards for drafting, revising, and grading outbound PM content with concise, active, defensible language. Use when Codex must produce or evaluate technical marketing content (summaries, launch messaging, web copy, narratives, or AI-generated drafts) and enforce Lanham-style prose tightening, lard-factor reduction, and layered quality checks.

Revise user-supplied prose exactly as Richard Lanham would in Revising Prose, using the Paramedic Method, Lard Factor cuts, CMOS-compliant American English, Oxford commas, and no contractions. Trigger whenever the user wants grammar-tight, persuasive rewrites that strip jargon and official-style bloat while staying truthful and reader-relevant.

Clean, trim, security-scan, validate, and optionally repackage ChatGPT/Codex skill folders or skill ZIP archives. Use when the user asks to clean up a skill folder, trim unused skill resources, remove redundant skill files, reduce skill package size, scan a skill for secrets, IPs, keys, credentials, hardcoded paths, or validate that cleanup did not break the skill.

Capture reproducible screenshots of web pages for documentation, QA, bug reports, baselines, and before/after comparisons. Use when Codex needs browser-rendered desktop/mobile/full-page/element screenshots from one or more URLs, including authenticated flows. Prefer Playwright automation and Chrome DevTools MCP tools; fall back to OS-level screenshot capture only when browser tools cannot capture the target.

Validate LiveLabs workshops end-to-end, including folder/manifest structure, Markdown formatting, and Richard Lanham language grading. Use when asked to audit a workshop root and emit VALIDATION-RESULT.md with file-by-file ratings.