The Agent Skills Directory

[PERSISTENCE]: The skill includes instructions to install a persistent background task on the host system using the command krankie cron install --hour 6, which sets up a daily cron job.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of an external package krankie globally via bun install -g krankie. This package is not from a trusted organization or well-known service, and it does not match the identified author context (jordancoin), presenting a supply chain risk.
[COMMAND_EXECUTION]: The workflow relies on the execution of several shell commands, including python scripts/validate_metadata.py and various krankie CLI tools for app and keyword management, which operate on the local environment.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted user-supplied data (app descriptions and features) and has the capability to execute shell commands.
Ingestion points: User-provided app details and optimization queries defined in the Core Workflow of SKILL.md.
Boundary markers: None identified; there are no instructions to the agent to treat user data as untrusted or to use delimiters.
Capability inventory: The skill can execute python and bun commands and manage a local SQLite database and cron jobs.
Sanitization: No evidence of input validation or sanitization before data is passed to the metadata validation or ranking tools.

app-store-aso