Security audit
Security checks for vulnerabilities and agentic risk
The skill is presented as a dotfiles backup and sync tool, but the inspected artifacts implement a local plaintext logging utility instead.
Review this carefully before installing. It does not appear to exfiltrate data or install a backdoor, but it also does not perform real dotfile backup, sync, restore, or version tracking. Treat it as a local plaintext logbook and avoid entering secrets, tokens, private config contents, or anything you would not want retained under ~/.local/share/dotfiles.
The skill is presented as a dotfiles backup/sync/versioning tool, but the documented behavior is actually a generic local logging and monitoring toolkit. This mismatch can mislead users or downstream agents into invoking it for backup or restore tasks that it does not perform, causing false assumptions about data protection, recovery, or system-state changes.
The implementation materially contradicts the declared purpose of the skill. A user invoking a 'Dotfiles' backup/sync tool would instead get a generic data-collection and logging utility with unrelated sysops-style commands, which is dangerous because deceptive capability mismatches erode informed consent and can trick users into supplying sensitive configuration data under false pretenses.
The advertised high-trust operations 'backup' and 'restore' do not back up or restore anything; they only persist arbitrary user input to log files. In context, this is especially dangerous because users may paste file paths, secrets, or configuration content expecting recovery actions, but the tool silently stores that data instead.
The manifest advertises backup and sync of dotfiles, while the body documents a general-purpose auditing/logging utility. In an agent setting, semantic mismatch is dangerous because tools are selected based on descriptions, so an agent may choose this skill expecting file protection or restoration and silently fail to achieve the intended security or operational outcome.
The branding and top-level framing suggest a dotfiles management skill, but the body immediately describes a sysops monitoring and reporting toolkit. This contradiction increases the chance of accidental misuse, especially by automated systems that may trust the title and early metadata more than the full body content.
The documented commands include dotfiles cleanup <input> and dotfiles restore <input>, and the usage section frames the skill as maintaining system configurations. Even though the table says these commands 'record' operations, the surrounding description suggests operational system-maintenance behavior, but no warning is provided about potential impact to files, configurations, or system integrity.
Labeling the script as a 'sysops tool' directly conflicts with the published dotfiles use case and is a deception/integrity issue. In a skill ecosystem, such contradictions are risky because users and reviewers may authorize execution based on the manifest while the code advertises a different operational role.
The overall design centers on collecting arbitrary user input, persisting it across commands, and later displaying or exporting it in plaintext. Given the declared dotfiles purpose, this context makes the behavior more dangerous because it can plausibly harvest sensitive configuration content under the guise of benign backup/sync functionality.
The help text markets the binary as a 'sysops toolkit' instead of a dotfiles manager, reinforcing the mismatch between claimed and actual behavior. This inconsistency increases danger in context because users may still trust the skill metadata and use sensitive backup/restore workflows that the tool does not perform.
The breadth of unrelated operational commands is inconsistent with a dotfiles-management skill and suggests the package is serving a different purpose than declared. While not directly code-execution dangerous, this hidden expansion of scope increases the likelihood of misuse, covert data collection, and user confusion about what the tool actually does.
The export feature writes accumulated logs into plain files without warning that previously captured user inputs will be copied and concentrated into new artifacts. This increases exposure because sensitive values may proliferate into additional files and formats that are easier to share, leak, or index.
The script persistently stores arbitrary user-supplied input without warning, consent, minimization, or sensitivity guidance. In a dotfiles context, users are especially likely to provide paths, hostnames, tokens, or configuration fragments, making silent retention a meaningful privacy and security risk.
The skill stores persistent per-command logs, history, and exports under a local data directory, but the user-facing guidance does not clearly warn about this persistence. Users may enter sensitive configuration details, hostnames, incident notes, or secrets assuming ephemeral processing, leading to unintended local data retention and later exposure.
No suspicious patterns detected.
