nature-reader Eval
==================

Status: review
Score: 82/100
Risk: medium
Decision: manual_review
Policy: review
Reason: Review the audit page, then allow agent install in a sandboxed workflow.

Install:
npx skills add Yuan1z0825/nature-skills --skill nature-reader

Required checks:
- PASS Task fit: Task wording matches this skill metadata.
- PASS Install path: Install handoff is available.
- PASS Install command safety: standard package or runtime install path
- WARN Trust score: Good trust signals with a few areas worth checking before rollout.
- PASS Audit score: Safe to try
- WARN Agent safety gate: Good audit and safety signals with no high-risk permission hints in public metadata.
- PASS License clarity: Apache-2.0
- WARN Permission surface: filesystem or document access, database access

Warnings:
- Trust score: Good trust signals with a few areas worth checking before rollout.
- Agent safety gate: Good audit and safety signals with no high-risk permission hints in public metadata.
- Permission surface: filesystem or document access, database access
- No critical security issues found. The skill reads local files and may download content from external sources (DOI/arXiv), but no destructive or exfiltration behavior is evident.
- The skill relies on a static/dynamic fragment system that requires careful file management; if any referenced file is missing, the skill may fail silently.

Validation plan:
1. Inspect repository, README/SKILL.md, license, and recent commits before production use.
2. Install in an isolated workspace or sandbox with no production secrets available.
3. Run the smallest representative task and record files touched, commands run, network access, and outputs.
4. Compare the selected skill against at least one alternative when the eval status is review or failed.
5. Promote only after the agent reports a successful verification result and unresolved warnings are accepted.

Do not use when:
- teams that need a vendor-supported SLA
- production agents without a repository review
- No critical security issues found. The skill reads local files and may download content from external sources (DOI/arXiv), but no destructive or exfiltration behavior is evident.
- No OpenAgentSkill engagement data yet
- The skill relies on a static/dynamic fragment system that requires careful file management; if any referenced file is missing, the skill may fail silently.
- Production credentials, payments, or irreversible account changes without explicit human review
- Sensitive private data before reviewing repository code, license, and permission surface
- Automatic installation in a production workspace

URLs:
- Skill: https://www.openagentskill.com/skills/yuan1z0825-nature-reader
- Audit: https://www.openagentskill.com/skills/yuan1z0825-nature-reader/audit
- JSON: https://www.openagentskill.com/api/agent/evals?slug=yuan1z0825-nature-reader