Security audit
Security checks for vulnerabilities and agentic risk
This is a simple local CSV preview/filter helper, with documentation inconsistencies but no evidence of hidden behavior, exfiltration, persistence, or destructive actions.
Install only if you want a local command-line CSV preview/filter tool. Treat the permission metadata as inaccurate: it reads local files, and the documentation should be cleaned up before using it in strict approval workflows. Avoid passing sensitive CSVs unless you are comfortable with their contents being printed to the agent or terminal output.
The manifest claims filesystem: false, but the skill explicitly requires a CSV file input and its documentation describes reading that file and potentially writing output. This creates a trust-boundary mismatch: policy engines or users may approve the skill under the false assumption that it cannot access files, when in practice its intended operation depends on filesystem interaction.
For markdown files, vague activation or invocation descriptions should be flagged when they are ambiguous or overly broad. The phrase "daily file/text chores" lacks clear scope or constraints and could overlap with many common user intents, making unintended invocation more likely.
The inline documentation says results may be emitted to a target file, directly contradicting the manifest's declaration of no filesystem capability. Such inconsistencies can cause unsafe approvals, incorrect sandboxing, or misuse by downstream agents that rely on the manifest for capability enforcement.
Describing optional file output without warning about file creation or overwrite risk can lead users or agents to unintentionally modify local data. In a file-processing skill, even non-malicious writes can damage workflows if outputs overwrite important files or occur in unexpected locations.
The title and Chinese description describe a narrow-purpose CSV row/column preview tool, while the English line says 'One command for daily file/text chores,' which implies a broader general-purpose utility. This is a documentation-level contradiction about intended scope, not merely missing detail.
This file’s user-facing natural-language content is entirely in Chinese, including the skill index and capability description. Under the stated policy, forcing a specific language without offering user choice or documenting a justified locale constraint is a natural-language policy violation.
The argument parser description and option help text are written entirely in Chinese, and later runtime messages also use Chinese-only output. This imposes a specific language on all users without any documented locale choice or opt-in, which matches the language-policy violation criteria.
