The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the ls command across multiple local directories, including home folder paths such as ~/.claude/skills/, ~/.agent/skills/, and ~/.agents/skills/. This is used to determine which companion skills are currently installed.
[EXTERNAL_DOWNLOADS]: In both the README and the SKILL.md instructions, the agent is directed to suggest the installation of companion skills from an external GitHub repository (AbsolutelySkilled/AbsolutelySkilled) which is not a verified or trusted vendor.
[REMOTE_CODE_EXECUTION]: The skill promotes the use of npx skills add AbsolutelySkilled/AbsolutelySkilled, a command that downloads and executes code from an external third-party repository at runtime.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8). It is designed to process external, untrusted keyword lists and content data while possessing the capability to execute shell commands and perform external downloads.
Ingestion points: User-provided keyword lists and seed topics (processed in SKILL.md).
Boundary markers: Absent. There are no explicit instructions to the agent to ignore embedded instructions within the keyword data.
Capability inventory: Shell command execution (ls) and remote installation capabilities (npx) defined in SKILL.md.
Sanitization: Absent. The skill does not provide methods for validating or escaping the user-provided data before processing.

keyword-research