
from vibe-research-skill20
Structure and maintain research projects with bi-directional sync between code, paper.md, and developer docs to keep experiments, documentation, and academic wr
Vibe Research provides a reproducible project layout and operational rules that let an AI agent manage research projects: it enforces bi-directional sync between source code, paper.md, and doc/ so documentation, experimental code, and academic write-ups stay consistent. It includes migration guidance, initialization scripts, and document maintenance practices.
Use this skill when starting or migrating a research project to a standardized structure, when you want the agent to maintain a running research log and docs, or whenever edits to code, paper, or docs must be kept in sync (e.g., preparing an academic paper or reorganizing experimental code). It is also useful for incremental migrations and audit workflows.
Friendly to general-purpose coding and assistant agents (Copilot/Codex/Code-focused models) and workflow automation assistants that can run shell scripts and edit repo files.
Vibe Research is a research project scaffolding skill that sets up directory structures and template files for bi-directional sync between code, paper.md, and developer docs. The init script creates directories successfully but fails when trying to copy template files because those resources aren't included in the scripts bundle — only init.sh was provided. SKILL.md is well-written with clear triggers, migration paths, and coding guidelines, but the skill is incomplete without the template resources.
template files (resources/templates/agents-rules.md, resources/templates/paper.md) not bundled in scripts dictClean script with no security concerns. Uses set -e, proper quoting, and idempotent file creation (checks existence before writing). The main issue is operational: template resources referenced by the init script aren't bundled, causing the script to fail. The 7-item directory rule is an interesting UX choice grounded in cognitive science. SKILL.md documentation quality is good with clear sections and examples.