ClawHub

AI Agent Helper

v1.0.0

AI Agent 設定同優化助手 - Prompt Engineering、Task Decomposition、Agent Loop設計

⭐ 4· 7.5k·115 current·122 all-time
by@katrina-jpg
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name, description and SKILL.md content all describe prompt engineering, task decomposition and agent loop design. There are no unrelated requirements (no env vars, binaries, or installs) that would be disproportionate to this stated purpose.
✓
Instruction Scope
SKILL.md contains guidance and examples for prompts, few-shot examples, output parsing and error handling. It does not instruct the agent to read files, access environment variables, call external endpoints, or collect/transmit data outside the agent—so the instructions stay within scope.
✓
Install Mechanism
No install specification or code is provided (instruction-only). This minimizes on-disk code execution risk and is appropriate for a prompt-engineering helper.
✓
Credentials
The skill requests no environment variables, credentials, or config paths. That is proportionate for a skill whose role is to provide prompt templates and design advice.
✓
Persistence & Privilege
always is false and the skill is user-invocable. disable-model-invocation is false (the platform default) but that alone is not a red flag for this benign, instruction-only skill.
Assessment
This skill appears coherent and low-risk because it's just an instruction file for prompt/agent design and asks for no installs or secrets. Before installing: 1) verify the publisher/source if you need provenance; 2) avoid pasting secrets or private data into prompts or examples provided to the skill; 3) if you enable autonomous agent invocation, limit the agent's tool access and scopes so it cannot reach external services or exfiltrate data; and 4) treat outputs as guidance—review any generated agent code or tooling instructions before running them in your environment.

Like a lobster shell, security has layers — review code before you run it.

latestvk9767agfs746q3v6yxhna6rs1981ejbz

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments