ClawHub

PPTAgent 协作助手

v1.0.0

Use when working in the PPTAgent or BotSlide repository to add features, fix bugs, or adjust slide generation, editing, export, prompt rules, project version...

⭐ 0· 139·1 current·1 all-time
byTonye@mrguanglei
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 target the PPTAgent/BotSlide codebase and the SKILL.md only references repository files, validation commands, and project-specific rules relevant to slide generation, preview, export, and versioning. There are no unrelated credential or platform requirements.
ℹ
Instruction Scope
Instructions direct the agent to read specific repo paths and docs, follow a debugging/validation checklist, and run repo-local verification commands (docker-compose, pnpm, pip, playwright, python smoke tests). These are reasonable for codebase collaboration, but they do imply executing build/test steps and starting services which may run arbitrary code from the repository or pull packages from registries.
ℹ
Install Mechanism
No install spec is declared (instruction-only), which minimizes persistent risk. However the included reference docs recommend commands that will fetch/install packages (pnpm, pip, playwright, docker images). Those network installs are expected for repo work but are not managed by the skill metadata.
✓
Credentials
The skill declares no required env vars, no credentials, and the instructions do not request secrets or unrelated system configuration. Access to repo files and running local verification commands is proportionate to the stated purpose.
✓
Persistence & Privilege
Skill is instruction-only, has no install step, and always:false. It does not request persistent presence or elevated agent-wide privileges.
Assessment
This skill appears coherent for collaborating on the PPTAgent/BotSlide repository: it points to repo files and suggests reasonable validation steps. Before running recommended commands, review the repository code and CI/test scripts so you know what will execute; run build/tests in an isolated environment (CI, container, or developer machine) because pnpm/pip/playwright/docker commands will download and run third-party packages. Confirm the repo contains no secrets or unexpected scripts (startup hooks, postinstall scripts) you don't want executed. If you only want read-only analysis, avoid running the install/start commands and ask for static guidance instead.

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

latestvk97034vyyzn2gzjaj6zy0ffhxs833675

License

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

Comments