Drawing Cleaner
AdvisoryAudited by Static analysis on May 8, 2026.
Overview
No suspicious patterns detected.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Running the skill will execute local code that reads the chosen Markdown file and creates or overwrites an output Markdown file.
The skill asks the agent/user to run a local Python script. This is expected for the tool’s purpose, and the provided script shows text processing and local file output only.
python scripts/clean_drawing.py [分栏提取_XXX.md] --output [清洗结果_XXX.md]
Use an explicit input and output path, and review the generated Markdown before using it in later workflow steps.
If invoked without a file path, the skill may clean the wrong Markdown file and write an output based on it.
If no input file is provided, the script automatically selects the first matching Markdown file in the current directory. This is convenient but could process an unintended file.
candidates = glob.glob("分栏提取_*.md") + glob.glob("*.md") ... input_path = candidates[0]Always provide the intended input file path rather than relying on automatic file selection.
Future runs could behave differently if the skill’s cleaning rules are edited.
The skill contemplates persistent updates to its filtering/classification rules. The wording ties this to user confirmation and keeps it scoped to regex rules.
如用户确认有漏判，记录到本 Skill 的 `NOISE_PATTERNS` 或 `CATEGORY_RULES` 中进行迭代。
Only allow rule changes you understand, keep a copy of the original script, and test changes on sample drawings before using them broadly.