writER
v1.0.1模仿 ER 个人写作风格的 skill。将口语、素材或草稿改写为富有诗意与哲思的散文/情书体文章，擅长意象营造与情感表达。
⭐ 0· 32·0 current·0 all-time
byEnrong Zheng@jewelroam
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (poetic/romantic prose style conversion) match the provided files (SKILL.md and references). The skill requires no credentials or binaries and does not attempt to perform unrelated system or network actions.
Instruction Scope
SKILL.md contains detailed, prescriptive transformation rules and examples limited to rewriting user-provided text. It does not instruct the agent to read system files, access environment variables, or transmit data to third-party endpoints as part of runtime instructions.
Install Mechanism
There is no formal install spec in the registry (instruction-only skill), which is low risk. README suggests manual copy or using a 'clawhub install' URL (https://clawhub.ai/jewelroam/er-writer-skill) — this is informational but means an agent could be instructed to fetch/install from that URL. The URL is not a shortened/personal IP, but users should verify it before allowing remote install.
Credentials
The skill declares no required environment variables, credentials, or config paths. SKILL.md does not reference secrets or external service tokens.
Persistence & Privilege
Skill flags are default (always:false, agent-invocable:true). The skill does not request persistent system privileges or modifications to other skills; nothing indicates it would alter global settings.
Assessment
This skill is an instruction-only style guide for rewriting text into a specific poetic/romantic voice and appears internally consistent. Before installing or asking an agent to install it: (1) inspect the files yourself (they are small and human-readable) to confirm you are comfortable with the content; (2) do not paste highly sensitive personal data into the skill—it will be sent to whatever model/service your agent uses; (3) if you permit the agent to run 'clawhub install' from the README URL, verify that URL/repository so you don't install untrusted code; (4) be aware of ethical/copyright considerations when asking the model to mimic a named person's distinctive style.Like a lobster shell, security has layers — review code before you run it.
latestvk97524kt2dx8mptah0g7qvkwwx8451y0
License
MIT-0
Free to use, modify, and redistribute. No attribution required.