nexu-io/codex-slides-research
Compatibility entry point for researching a current topic into a source-backed Markdown brief and turning it into a Codex Slides outline or rendered deck. Use for market, competitive, policy, technical, academic, or other presentation requests that need current facts, figures, comparisons, or citations.
npx skills add https://github.com/nexu-io/codex-slides --skill codex-slides-research
The canonical codex-slides sibling skill owns the complete workflow. Use
deep_research when MCP is available; otherwise run CLI command research at
../codex-slides/scripts/codex-slides.mjs. Preserve its Markdown and citations
so the user can review or edit the evidence before slide generation.
For the default interactive workflow, first open the Codex Slides home screen
in the Codex Browser, submit the research request there, and keep the created
project visible while its research, outline, style, and render checkpoints
advance. Use the lower-level sequence below only for CLI automation, recovery,
or a user-requested headless workflow:
list_scenarios / CLI scenarios andpass the selected scenarioId into deep_research / CLI research with
one to four rounds. This carries the scenario's evidence contract into the
search brief.
create_outline / CLI outline with researchDoc set to the reviewed brief.revise_outline / CLI revise-outline when needed.render_deck / CLI render only after the outline is accepted.open_codex_slides / CLI open and navigate the returned projectresource or URL in the in-editor Browser for visual verification.
For an explicitly requested unattended fast path, call create_deck with
research: true. Do not claim a brief is source-backed when its returned
Markdown has no source links.
For a durable outline revision or render that should survive navigation, use
start_project_run, then retain and track its runId while the Browser remains
on the matching checkpoint.
Take nexu-io/codex-slides-research from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.