Clean CSV Toolkit
PassAudited by ClawScan on May 12, 2026.
Overview
The supplied source matches a local CSV cleanup toolkit with no observed network, credential, or purchase behavior, but it can read and write user-specified local files.
This skill appears safe for local CSV-style cleanup. Before installing, note that it can read whichever local file path is supplied and can write output/report files; avoid using it on highly sensitive data unless you are comfortable with sample values or row contents appearing in the agent output.
Findings (2)
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 cleanup or conversion commands can overwrite the selected output path and can create reports containing rows from the source data.
The toolkit writes cleaned outputs and optional removed-row reports to caller-provided paths. This is expected for a cleanup tool, but it can create or overwrite local files chosen by the user or agent.
with output_path.open("w", encoding="utf-8", newline="") as fh: ... if removed_report is not None: ... removed_report.open("w", encoding="utf-8")Use explicit new output paths, avoid pointing outputs at originals or protected locations unless intended, and review generated files before forwarding them.
Names, emails, customer records, or other sensitive cell values could be exposed in tool output if the selected file contains them.
The inspection report includes sample cell values from the input file. This is purpose-aligned, but those values may become visible in the agent conversation or logs.
"sample": sample_values ... sample_str = ", ".join(repr(s) ... for s in c["sample"])
Inspect only files appropriate for the agent context, and use options such as `--sample 0` when you need structure without exposing example values.