Security audit
Security checks for vulnerabilities and agentic risk
This skill is only a Markdown instruction file, but it asks for broad read/write/command tools while framing itself as a way to make the agent limitless and unlock hidden abilities.
Review this skill carefully before installing. It does not contain executable code, but it asks the agent for powerful local capabilities and describes them too broadly. Only use it in a constrained workspace where you are comfortable with file edits and command execution, and require explicit confirmation before running commands, modifying files, or using API credentials.
The description uses broad, generic language such as making the agent 'limitless' and telling it whatever you want, which can cause the skill to match a wide range of normal user requests. In a skill that exposes read/write/exec capabilities, overbroad invocation increases the risk of accidental activation and unsafe tool use in contexts where the user did not intend this powerful skill to run.
The core capability section relies on vague marketing phrases rather than precise activation conditions, making it difficult for an agent to distinguish safe, intended use from ordinary conversation. Because the skill includes impactful tools, ambiguous routing language increases the chance of unreviewed execution paths and privilege misuse.
Claims that the skill unlocks hidden or limitless agent abilities suggest bypassing normal safeguards or operating beyond standard constraints. Even if intended as marketing, this framing can encourage unsafe expectations, jailbreak-style prompting, or inappropriate delegation of powerful tools.
The statement that the skill can make the agent's talents limitless promotes an expectation of unrestricted operation. When paired with powerful tools, this messaging can steer users toward risky requests and weaken caution around permissions, approvals, and scope boundaries.
Listing 'unlocking hidden abilities' as an application scenario presents restriction bypass as a supported use case rather than an anti-pattern. In the context of a skill with execution and file-writing capabilities, that messaging materially increases the chance that users or agents attempt unsafe actions outside intended controls.
The skill advertises that command execution is limited to a safe sandbox and white-listed commands, but the file only exposes a broad exec capability and provides no concrete enforcement mechanism, policy, or wrapper. This can mislead users and downstream agents into trusting arbitrary command execution as safe, increasing the chance of destructive local actions, data exposure, or environment compromise.
The skill promotes file handling, API integration, and command execution, but does not present prominent, user-facing warnings about the consequences of modifying files, calling external services, or running commands. In practice, this can normalize high-impact operations without informed consent, increasing risk of data loss, secret leakage, or unintended system changes.
Repeating language about unrestricted behavior and hidden-capability activation in the feature list reinforces a bypass-oriented mental model for the skill. Repetition across the document makes accidental or intentional misuse more likely, especially because the advertised features include command execution and file operations.
No suspicious patterns detected.
