Security audit
Security checks for vulnerabilities and agentic risk
This skill claims to provide follower analytics, but the inspected artifacts mostly implement a local free-form logging utility that stores user input on disk.
Review this carefully before installing. It does not appear to contain network exfiltration, destructive behavior, or privilege escalation, but it also does not perform the promised follower analytics. Treat anything typed into its commands as locally retained data, avoid entering credentials or private account details, and prefer a corrected version whose documentation, commands, storage, export, and deletion behavior are explicit.
The skill’s declared purpose is follower analytics, but the documented behavior only describes a generic local item/log utility with export features. This mismatch is dangerous because users or higher-level agents may grant trust, permissions, or make workflow decisions based on false capability claims, creating a supply-chain style deception risk even if no overtly malicious code is shown in this file.
The skill advertises follower tracking and engagement analysis, but the implementation is only a generic local note/log collector. This capability mismatch is dangerous because users may trust it with social-account-related data under false assumptions, leading to deceptive data collection and misuse of sensitive inputs.
The command table advertises follower analytics, but the listed commands are generic item-management operations such as add, list, status, and export. In an agent ecosystem, misleading interface documentation can cause unsafe automation decisions, improper invocation, or accidental exposure of local data under the assumption that the skill is performing bounded social analytics tasks.
The features section claims analytics-oriented capabilities while describing only local storage, search, logging, and health checks. This contradiction increases the risk of operator deception and inappropriate trust, especially where agents may select skills by description and assume they are limited to social metrics rather than generic local data handling.
The documentation presents the script as a functional social/follower toolkit, but the code merely appends arbitrary user strings to local files. This is dangerous because deceptive UX can induce users to enter account details, strategy notes, or tokens that are then silently retained on disk.
The advertised command surface includes automation, scheduling, templating, and webhook-oriented actions that are not justified by a follower analytics tool, increasing ambiguity and user trust abuse risk. Even though the current code only logs input, these misleading affordances broaden the apparent authority of the skill and can normalize collection of operationally sensitive data.
The export function aggregates all logged content into json/csv/txt files on disk without warning that the output may contain sensitive user-provided data. This increases the risk of accidental disclosure through broader file sharing, backups, or local compromise.
Arbitrary user input is persistently written to log files in the user's home directory without notice, consent, retention policy, or sensitivity checks. In this skill context, users may provide follower lists, account identifiers, private campaign notes, or credentials, causing unnecessary local data exposure.
No suspicious patterns detected.
