Draft, revise, or critique a primary-research manuscript for a specifically chosen Nature or Nature Portfolio journal, or deliberately develop that broad-reader scientific style. Ground claims in the supplied results and calibrate journal requirements separately. Do not impose this journal style on generic documentation, reviews, press releases, or unrelated writing.
Use this Skill: https://skilld.dev/gh/tristanmanchester/agent-skills/nature-article-writer Nothing lands on disk. Nothing to clean up.
Edit a local copy. It keeps the author and licence.
≈171 tokens on demand. Your agent reads this file only when SKILL.md points to it.
Use one row per figure or conceptual unit.
| Figure | Scientific question | Core observation | Allowed claim | Main uncertainty / caveat | Key statistical details for legend | Methods detail the reader must know |
|---|---|---|---|---|---|---|
| Fig. 1 | ||||||
| Fig. 2 | ||||||
| Fig. 3 | ||||||
| Fig. 4 | ||||||
| Fig. 5 | ||||||
| Extended Data / Supplementary |
Source: SKILL.md on GitHub
Runs shell commands
skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.
skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.
The skill helps users draft scientific manuscripts for Nature Portfolio journals. It contains deceptive metadata claiming the author is OpenAI, which misrepresents its source. It also processes external drafts and notes without isolation boundaries, creating an indirect prompt injection surface, and directs the agent to execute shell commands to run provided Python diagnostic scripts.
No alerts
Risk: LOW · No issues
Signed by skilld at 3323bc9. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.
Last checked against GitHub 2 weeks ago.
{
"version": "3.0.0",
"reviewed": "2026-09-13"
}