Security audit
Security checks for vulnerabilities and agentic risk
This skill is a straightforward local data-visualization tool that reads a user-specified CSV, Excel, or JSON file and writes a chart output file, with no evidence of hidden persistence, credential access, exfiltration, or destructive behavior.
This skill is reasonable to install for local chart generation. Before using it on confidential data, remember that generated HTML chart files may contain the underlying data needed for interactivity, so treat outputs like the source dataset. For stronger supply-chain hygiene, prefer a locked environment with reviewed versions of pandas, plotly, and openpyxl.
The trigger phrases are broad enough to match many ordinary data-analysis or charting requests, which can cause the skill to activate outside its intended scope. Overbroad activation increases the chance an agent routes sensitive files or unrelated user tasks into this skill, expanding attack surface and creating unintended tool use.
The Chinese keywords are extremely generic terms for visualization and charting, so they are likely to collide with normal conversation and unrelated analytics requests. This can lead to accidental invocation in multilingual contexts, especially where the agent relies heavily on keyword matching.
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
Create visualization from data file.
python3 ~/.openclaw/workspace/autoplot/scripts/autoplot.py visualize <file> [options]
The dependency is specified with only a lower bound, so builds may resolve to different versions over time and can silently pick up vulnerable or incompatible releases. In a skill that automatically processes user-supplied data files, dependency drift increases supply-chain risk and makes security review and incident response harder.
pandas>=1.5.0
plotly>=5.0.0
openpyxl>=3.0.0
The manifest does not pin pandas, so it is impossible to verify whether deployed environments avoid versions associated with known advisories. While the cited CVE is disputed and not necessarily exploitable here, the lack of version control still creates uncertainty around whether vulnerable releases could be installed.
The unpinned plotly dependency allows non-reproducible installs and may introduce vulnerable or unexpected upstream changes without any modification to this repository. Because this skill generates output from arbitrary user data, predictable and reviewable dependency versions are important for reducing supply-chain exposure.
pandas>=1.5.0
plotly>=5.0.0
openpyxl>=3.0.0
Using openpyxl with only a minimum version means the environment may install any later release, making it impossible to guarantee that a safe version is used. This is more relevant here because the skill explicitly ingests Excel files, so the package is on a direct parsing path for attacker-controlled input.
pandas>=1.5.0
plotly>=5.0.0
openpyxl>=3.0.0
openpyxl has historical XXE-related advisories, and because this skill processes Excel files from users, an unpinned version leaves uncertainty about whether a vulnerable parser could be installed. If an affected release were used, crafted spreadsheet content could potentially trigger unsafe XML entity handling during file parsing.
No suspicious patterns detected.