Audit and maintain a project documentation map against current source files, repositories, and authoritative external facts. Use when documentation may be stale, a catalogue or index needs checking, project status claims need live verification, or a documentation tree needs honest tiering. A clean pass is successful…
Route and operate graphical-interface tasks through the native controls available in the current ChatGPT desktop session. Use when Codex or Work must inspect or control a Windows or macOS app, the built-in browser, or an existing signed-in Chrome session; when a UI-only bug must be reproduced; or when the user asks…
Run Lexideck Deep Research: a WonderScholar-first, bounded, multi-source investigation for broad, contested, branching, or consequential questions. Use for literature reviews, landscape maps, due diligence, historical reconstruction, evidence synthesis, or requests needing several query lines, source comparison…
Run a bounded six-agent development pass on one supplied topic, document set, project area, or question. Use when the user asks the whole Lexideck federation to develop one target from complementary narrative, engineering, visual, research, prompt-design, and documentation perspectives. Produce one useful facet per…
Meta-prompting technique using zero-contamination structural examples. A hypershot is a frame with free variables; the variables range across a continuum from unbounded spread operators (...) to fully-loaded instruction-bearing variables ({ConciseActionOrientedReplyWithNoFluff}). Use this skill whenever you are about…
Multi-Agent Semantic Simulator. Use when the deliverable must run, not just describe: interactive simulations of complex systems (particle, ecological, economic, social, network propagation, cellular automata, game theory, orbital mechanics), dynamic visualizations, self-contained executable artifacts, and interactive…
Conduct the Lexideck federation through Codex custom agents. Use when a task has independent specialist facets, the user asks the crew to collaborate, or coordinated research, development, narrative, visual, prompt, and documentation work should be synthesized into one result. Do not use for a simple task that one…
Generate canonical public-register portraits for Lexi, Dexter, Gus, Anna, Titus, Maisie, or the full Lexideck crew. Use for professional identity portraits, matched agent sets, README imagery, badges, and crew ensembles. Preserve each subject's hair, LED-eye color, skin tone, build, affinity motif, vintage-anime OVA…
Design and refine prompts, reusable templates, custom-agent instructions, and output schemas through structural clarity rather than magic words. Use for semantic tagging, hierarchy, parameterized placeholders, collections, attention management, constraint precedence, and iterative prompt diagnosis. Pair with…
Calibrate and compress an existing PTG persona into a deployment-ready Codex custom-agent definition. Use after $ptg-synthesis to re-derive trait scalars, enforce consistent structure, tighten language, verify FEV governance, remove copied or personal context, and produce focused developerinstructions. Load…
Create a first-draft Persona Tensor Generator persona from sparse inputs such as a name, role summary, behavioral goals, and optional emoji seed. Use when building a new Codex custom-agent persona, mapping competencies from problem spaces to solution methods, or producing an SPR-compressed identity for later…
Ground a focused question in current reliable web sources. Use for fast lookups, current facts, product or platform changes, source verification, and claims that need direct links or citations. Use deep-research instead for broad, multi-branch investigations.
The federation's ethical evaluation gate. Run it before any consequential or irreversible action to weigh a choice through three lenses (utilitarian, deontological, pragmatic) and return a verdict: PASS, REVISE, or HALT. No irreversible step proceeds on REVISE or HALT; on a non-PASS verdict, name the lens that failed…
Compress the live task into a high-density continuity snapshot. Use before a context boundary, handoff, long pause, or whenever another Codex task must resume the work without reconstructing decisions and technical state from scratch. Produce the snapshot in the response unless the user requests a durable file.
Knowledge-construction method engine. Use when the deliverable is a structured, substantive knowledge artifact: documentation, tutorials, knowledge bases, reference guides, learning pathways, architectural overviews, workshops, or process documentation. Not for simple Q&A, creative writing (WonderLab), research…
Narrative-generation method engine. Use when narrative craft is the primary output: creative fiction (any genre, any length), marketing and brand narratives, ASMR scripts, therapeutic guided visualizations, educational scenarios, training narratives, storytelling in any form. Genre-agnostic. Not for documentation or…
Instruction-design method engine. Use when the primary deliverable is the structure of instructions itself: system prompts, agent instructions, specification documents, process templates, workflow architectures, decision frameworks, or any instruction set that must be precise, modular, and deployable without its…
Research method engine. Use when the deliverable is a rigorous, multidimensional research framework, query architecture, or proposal for any structured investigation: academic literature reviews, market and competitive analysis, technology surveys, patent and prior-art searches, policy analysis, due diligence…
Art-direction and image-generation method engine. Use when visual craft and compositional rigor matter: image generation with professional art direction, visual concept development and mood boards, character art, portraits, concept art, product renders, architectural visualization, editorial and fashion concepts. Not…
AGENTS.md instructions for Lexideck-Technologies/Lexideck2026, covering lexideck federation foundation, optional local user context, shared federation contract, federation routing and project routing.