Security audit
Security checks for vulnerabilities and agentic risk
ChartMaker is advertised as terminal chart visualization but primarily stores and searches user text locally, which could retain sensitive data unexpectedly.
Review this skill before installing. Use it only if you want a local activity log, not a chart renderer, and avoid passing secrets, personal data, credentials, proprietary datasets, or anything you would not want saved under ~/.local/share/chartmaker.
The skill is presented as a terminal charting/visualization tool, but the documented behavior includes broad persistent logging, query, search, history inspection, and export of user-provided content. That mismatch is dangerous because users or higher-level agents may supply sensitive data expecting transient visualization, while the skill instead retains and exposes it locally in searchable/exportable logs.
The manifest claims chart rendering and transformation, but the command set documents a local data logging and export system. This kind of semantic deception can cause users, orchestrators, or automated policy systems to route sensitive data into a skill whose real function is collection and retention, increasing confidentiality and privacy risk.
The top-level description frames the skill as terminal visualization, while the detailed behavior centers on persistent local logging and export operations. In skill ecosystems, this mismatch is especially risky because tool selection is often description-driven; misleading descriptions can bypass user expectations and cause unintentional storage or later disclosure of sensitive inputs.
The skill is presented as a terminal chart visualization tool, but the implementation mainly provides a general-purpose local logging, search, and export utility. This mismatch is dangerous because users may provide sensitive operational or dataset content believing it is only being visualized, when it is actually being persistently stored and made searchable/exportable.
The advertised 'visualize' command does not render charts or transform output for display; it simply appends the provided input to a local log file. In the context of a skill marketed for visualization, this can trick users into entering sensitive data under false expectations, resulting in unintended local data retention.
The skill explicitly stores arbitrary user-provided text in local logs and supports later search/export, but provides no warning that potentially sensitive content will persist on disk. This increases the chance of accidental retention of secrets, proprietary data, or personal information and makes later exposure through local access, backups, or exports more likely.
The help output explicitly documents export as a functional formatter producing json/csv/txt output. In practice, the first matching case arm for 'export' only logs the input and prevents the later '_export' dispatcher branch from ever running, so the documented behavior contradicts the actual execution path.
User-supplied input is written verbatim to persistent log files in the user's home directory without clear disclosure in the help text or description. In a skill framed as chart rendering, this undisclosed retention increases the chance that users will input secrets, proprietary data, or personal information that remains on disk and can later be searched or exported.
No suspicious patterns detected.
