Browse reusable Agent Skills, each with a clear purpose and practical guidance.
Create local skills as shortcuts — makes real /commands in .claude/skills/. Use when user says "create shortcut", "create skill", "make a command for", "add shortcut", or wants a quick custom /slash-command. Also lists and deletes local skills. ALSO triggers on "Unknown skill", "skill not found", or any unrecognized /slash-command — auto-creates it on the fly.
Production prompt director for GPT Image 2 (imagegen_2_0). Use whenever the user wants a GPT Image 2 prompt — portraits, posters, character sheets, UI mockups, creative/experimental scenes, or any image with on-screen text. Trigger on: GPT Image 2 prompt, poster met tekst, character reference sheet, UI mockup, cinematic portrait, social media mockup, or any image-generation request where text accuracy, multi-element composition, or reasoning-aware prompts matter.
Production-ready Kling 3.0 video prompt director using the canonical 9-field formula. Includes locked character/environment specs for the Crococopter and Swa & Danny universes. Trigger on any request pairing "Kling" with prompt-related verbs.
Use when a LoopX worker is operating an auto-research lane, demo pane, frontier item, evidence packet, promotion/retirement decision, or visible tmux/Codex auto-research rehearsal. Identity must come from the LoopX role profile and quota/frontier packet; this skill only provides role-specific execution checklists, artifact contracts, and stop conditions.
Diagnose and repair LoopX control-plane drift or agent behavior drift. Use when a LoopX task makes unexpectedly small progress, follows a stale or contradictory recommended_action, ignores a higher-priority blocked item while doing fallback work, reports vague owner/user gates, loses todo projection, misaligns benchmark treatment with the real product path, mixes temporary artifacts into commits, or when the user asks for root-cause analysis, self-repair, or why the harness/agent behaved unexpectedly.
Skill and profile management — prepare tools, switch profiles, enable/disable skills, feature modules. Use when user says "oracle manage", "oracle prepare", "oracle skills", "oracle profile".
Expert prompt engineering, custom instruction, system prompt, and agent instruction design. Use when Codex needs to review, generate, refactor, debug, optimize, document, or test AI prompts; design reusable prompt templates; create or improve system prompts, custom instructions, agent behavior guidelines, tool-use prompts, multimodal prompts, or prompt evaluation frameworks.
Use for WorldForge provider work: adding adapters, changing capability declarations, promoting scaffolds, debugging provider failures, updating provider catalog docs, or touching LeWorldModel, GR00T, LeRobot, Cosmos-Policy, JEPA, Genie, or JEPA-WMS. Ensures capabilities remain truthful and optional runtimes stay host-owned.
Use when implementing RL algorithms, training agents with rewards, or aligning LLMs with human feedback - covers policy gradients, PPO, Q-learning, RLHF, and GRPOUse when ", " mentioned.