Adlaw Checker
v1.0.0检测文案中的广告法违禁词及虚假宣传内容，提供风险评估和修改建议，支持多平台合规检查。
⭐ 0· 48·1 current·1 all-time
by@eric060
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (广告法违禁词检测) match the provided implementation: static violation lists, platform rules, suggestion generation, batch/CLI interfaces. The SKILL.md's recommended Python deps (jieba, pandas) are reasonable for text processing, even though the visible portion of the script uses only core Python; overall capability and requested resources are proportional.
Instruction Scope
Runtime instructions stay within the domain of checking and transforming text: CLI, Python API, batch checks, optional file monitoring. The SKILL.md does not instruct the agent to read unrelated system files, access external endpoints, or exfiltrate data. One minor note: SKILL.md mentions monitoring a local file with a callback — that is within expected scope but the full implementation of monitoring/callbacks should be reviewed before running.
Install Mechanism
There is no automated install spec (instruction-only); installation is manual (pip install jieba pandas and copy folder). No downloads from arbitrary URLs or archive extraction are present in the package metadata. Risk from install mechanism is low.
Credentials
The skill does not request any environment variables, credentials, or config paths. All declared/used data are local inputs (text, files). This is proportionate to the stated purpose.
Persistence & Privilege
Skill flags are default (always: false, user-invocable: true) and it does not request permanent presence or system-wide configuration changes. No modification of other skills or elevated privileges are indicated.
Assessment
This skill appears coherent and local-only: it scans text for forbidden/overstated phrases and proposes replacements. Before installing or running: (1) inspect the remainder of scripts/checker.py (the provided file was truncated) to confirm there are no network calls, telemetry, or unsafe eval/exec usage; (2) run it in a sandbox or virtual environment and pin / audit the pip packages you install (jieba, pandas) from PyPI; (3) if you will use the monitor(callback) feature, review the callback flow so arbitrary code isn't executed unexpectedly; and (4) remember results are not legal advice—perform human/legal review for important documents.Like a lobster shell, security has layers — review code before you run it.
latestvk97e4gw84t64wg67vcb7e8sf3s84c2nx
License
MIT-0
Free to use, modify, and redistribute. No attribution required.