The Agent Skills Directory

[UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The SKILL.md file contains instructions to install the 'parallel-cli' tool by piping a script from 'https://parallel.ai/install.sh' directly into bash. This method of installation is insecure as it allows for the execution of arbitrary code from a remote source without verification or integrity checks.
[INDIRECT_PROMPT_INJECTION]: The skill's primary functionality involves searching and extracting content from external web URLs, which serves as an ingestion point for untrusted data. This content could contain malicious instructions designed to hijack the agent's logic.
Ingestion points: The 'parallel-cli search' and 'parallel-cli extract' commands in the reference files are used to pull arbitrary content from the web into the agent's context.
Boundary markers: There are no explicit instructions or delimiters used to warn the agent to ignore embedded instructions within the fetched web content.
Capability inventory: The agent has the ability to execute shell commands, perform further network requests, and write results to the local file system using the provided CLI tools.
Sanitization: The skill does not define any sanitization, filtering, or validation steps for the content retrieved from external URLs before it is processed by the AI.

parallel-web