Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed, gated workflow for creating one short image-to-video clip, with a minor installation supply-chain caution around unpinned npx commands.
Before installing, consider using a trusted or pinned version of the `skills` CLI instead of running unpinned `npx` commands directly. During use, expect the workflow to upload selected media/audio to Pruna-related services and potentially make paid generation calls only after the listed approval gates.
The skill instructs the agent/user to execute npx skills add ... without pinning the skills package itself to a specific version. Because npx resolves the package at runtime, a future compromised or breaking release could change behavior or execute unintended code during installation, creating a supply-chain risk.
This line again relies on unpinned npx skills, which fetches and runs the current package version at execution time. If the upstream package or dependency chain is compromised, the agent could install or run attacker-controlled code.
Using npx skills without an exact version leaves the installation path non-deterministic and vulnerable to supply-chain changes. In an agent skill context, these commands may be copied or automated, increasing the chance of executing an unexpected package version.
The command references an unpinned runtime-resolved package via npx, which can silently pull newer or malicious versions. That creates a realistic software supply-chain exposure rather than just a documentation issue.
Because the skill directs installation through unpinned npx skills, it trusts the latest package state at execution time. If exploited, this could lead to arbitrary code execution in the environment performing the install.
This command inherits the same supply-chain risk from unpinned npx skills: a mutable upstream package is executed on demand. In a skill file, such commands are especially risky because users may trust and run them without additional validation.
The skill references a package manager execution pattern that is not version-locked for the actual skills executable. This exposes consumers to malicious updates or dependency confusion-style issues if the package resolution path is subverted.
An unpinned npx skills command means the tool used to install other skills is itself mutable. If the CLI is compromised, every downstream install action becomes a vector for broader compromise.
This line uses the same unpinned npx execution model, which can execute whatever the registry currently serves. In a workflow handling uploads and paid API actions, compromise of the installer could have access to secrets, files, or billing-linked operations.
The full-suite install also uses unpinned npx skills, broadening the blast radius because one command bootstraps many capabilities. A compromised or altered CLI version could affect all installed skills at once.
The related-skills section repeats the same unpinned npx skills execution pattern. Even though it is documentation, users or agents may directly follow these commands, making the supply-chain risk operational.
This unpinned install command can resolve to different code over time, undermining reproducibility and trust. If a malicious version is published, it could execute during installation with the user's privileges.
Again, the command relies on mutable package resolution through npx skills. In an agent ecosystem, this is more dangerous because automated systems may execute documented commands without human scrutiny.
The same package bootstrap issue appears here: unpinned npx skills can silently change over time or be compromised upstream. That creates a practical path to arbitrary code execution during setup.
This final related-skill install command repeats the unpinned npx skills pattern. The risk is compounded by repetition because consumers are encouraged to use the same insecure bootstrap mechanism throughout the skill.
No suspicious patterns detected.
