The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation instructs the use of npx ai-social, which triggers the download of the ai-social package from the public npm registry at runtime.
[REMOTE_CODE_EXECUTION]: By utilizing npx, the skill executes code from a remote source that is not part of the trusted organization list. This allows for the execution of arbitrary logic contained within the external package.
[COMMAND_EXECUTION]: The skill documentation encourages the execution of shell commands using user-supplied file paths as arguments. This could lead to unintended command execution if the agent does not strictly validate the input file names or paths.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from local files (e.g., README.md) to generate social media content using an LLM.
Ingestion points: Processes content from arbitrary markdown or text files provided as command-line arguments.
Boundary markers: No delimited sections or instructions are provided to the agent to ignore potentially malicious content within the source files.
Capability inventory: The skill reads local file content, executes shell commands via npx, and requires an OPENAI_API_KEY environment variable for external API calls.
Sanitization: There is no evidence of sanitization, filtering, or validation of the input file content before it is processed by the external tool.

social-gen