多 Agent 项目协作工厂

v1.0.0

多 Agent 项目协作工厂 - 一键实现智能分工完成互联网项目

0· 55·0 current·0 all-time
byninetyhe@ninetyhe-90
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 (multi‑agent project builder) aligns with included assets: a simple project generator script and templates for configuration, README, and execution guide. Nothing in the bundle requests unrelated privileges or credentials.
Instruction Scope
SKILL.md stays on‑topic: it instructs how to generate a project, configure agents, and run within Hermes. Minor note: examples reference running the provided script from a specific ~/.hermes/skills/... path — this assumes installation layout but does not itself read or exfiltrate other files or secrets.
Install Mechanism
No install spec; this is instruction‑plus‑templates only. The included Python script is local, has no network calls, and only performs filesystem copy/create operations. No downloads or archive extraction are present.
Credentials
The skill declares no required environment variables, credentials, or config paths, and the code does not access environment secrets. That matches the described functionality.
Persistence & Privilege
always is false and the skill does not request persistent system privileges or modify other skills or global agent configuration. It only writes files within user‑created project directories.
Assessment
This skill appears coherent and low risk: it creates a project folder and copies template files. Before running: (1) inspect the templates and the short Python script (done here — it only writes files and asks before overwriting), (2) run it in a safe/work directory to avoid accidental overwrites, and (3) avoid placing secrets or API keys into agent custom_prompt fields or PROJECT_CONFIG.yaml. If you plan to let the agent run autonomously, be mindful not to grant it any external credentials or paste sensitive data into prompts.

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

latestvk971ddq1zr4qkwmfsjp5r36r7n84p9pm

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments