Create and review hooks for short-form creator content. Use when Codex is asked to write TikTok, Reels, Shorts, LinkedIn, X, or short-video hooks, opening lines, pattern interrupts, or retention-focused first seconds.
Design lightweight benchmarks for agent skills and skillsets. Use when Codex is asked to prove whether a skill improves outputs, define benchmark scenarios, compare baseline versus skill-enabled behavior, or create a skill evaluation plan.
Evaluate, forward-test, and improve Codex skills. Use when Codex is asked to test a skill, compare performance with and without a skill, improve a skillset, make skills stronger, validate skill usefulness, create acceptance tests for skills, or iterate skill instructions based on realistic scenarios.
Score agent skill outputs with a rubric. Use when Codex is asked to evaluate a skill output, compare outputs, apply a quality rubric, summarize reviewer scoring, or decide whether a skill output is good enough to promote.
Threat model agent skills, skillsets, and install paths. Use when Codex is asked to identify risks in a skill, define assets and trust boundaries, review permissions, or summarize security risks before installing or publishing a skill.
Review the context, token, tool, and process overhead added by agent skills. Use when Codex is asked whether a skill is too verbose, too broad, costly to invoke, duplicative, or worth its context/tool overhead.
Plan, create, export, and review social-ready visual assets such as profile pictures, banners, post images, article covers, thumbnails, carousel slides, short-video covers, and platform crop variants. Use when Codex is asked to make or audit social graphics, crop-safe images, profile-circle previews, banner previews…
Plan SQL analyses without assuming direct database access. Use when Codex is asked to design SQL queries, outline joins, filters, CTEs, aggregations, validation checks, or data quality steps for analytics and BI questions.
Review agent skill supply-chain risk. Use when Codex is asked to inspect third-party skill sources, dependencies, install commands, package managers, remote downloads, version pins, provenance, or update risks before installation or publication.
Prepare review-only synthetic character bibles for AI-assisted creator personas, mascots, recurring fictional hosts, virtual influencers, or synthetic spokescharacters. Use when Codex is asked to define a synthetic character's identity, voice, visuals, backstory, boundaries, disclosure rules, likeness and voice risks…
Build, audit, and refactor Tailwind CSS interfaces and design systems. Use when Codex is asked about Tailwind classes, component variants, utility consistency, tokens, theme variables, responsive patterns, class detection, dynamic class issues, production CSS size, design-system enforcement, or Tailwind UI polish.
Create YouTube and video thumbnail/title creative briefs. Use when Codex is asked to write title packs, thumbnail concepts, visual contrast notes, curiosity gaps, A/B test ideas, or title-thumbnail alignment checks.
Classify and prioritize customer support tickets. Use when Codex is asked to triage support requests, assign severity, identify customer impact, summarize ticket facts, route ownership, detect missing information, or prepare next support actions.
Write task-focused tutorials for developers, users, maintainers, or operators. Use when Codex is asked to create a walkthrough, quickstart, how-to, learning path, or step-by-step guide that leads to a verified outcome.
Generate or plan focused unit tests for functions, modules, components, and services. Use when Codex is asked to add unit tests, improve unit coverage, define test cases, create fixtures, or guard logic changes with small deterministic tests.
Review creator sponsorship usage-rights questions and checklist items. Use when Codex is asked to identify influencer usage rights, paid media usage, whitelisting, licensing, exclusivity, term length, territory, ownership, brand approvals, or rights questions for counsel or manager review.
Explain variance between actuals, targets, forecasts, plans, prior periods, or cohorts. Use when Codex is asked to write finance commentary, explain budget or revenue variance, compare KPI actuals to plan, summarize month-over-month changes, or prepare variance notes for review.
Prepare vendor negotiation briefs for owners and operators. Use when Codex is asked to prepare for vendor renegotiation, compare vendor options, draft negotiation talking points, review renewal tradeoffs, or create a vendor decision brief.
Write creator video scripts for short-form, long-form, sponsored, educational, story, and product content. Use when Codex is asked for a video script, shot-by-shot outline, voiceover, talking points, b-roll notes, or retention beats.
Audit web and app interfaces using visual evidence. Use when Codex is asked to critique UI/UX, polish a design, inspect a localhost app, compare screenshots, find visual hierarchy/layout/typography/color/spacing/state/responsiveness/accessibility issues, or provide annotated design feedback grounded in browser…
Write and review product requirements documents, feature specs, implementation briefs, and acceptance criteria. Use when Codex is asked to turn research, feedback, product decisions, or roadmap ideas into a PRD, product spec, feature brief, requirements doc, or implementation-ready product plan.
★not rated 1 8d agoA58 tokens
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: