The Agent Skills Directory

[PROMPT_INJECTION]: The skill interpolates user input directly into the prompt via the $ARGUMENTS variable without using delimiters or boundary markers. This creates an indirect prompt injection surface.
Ingestion points: The $ARGUMENTS variable in SKILL.md captures user questions.
Boundary markers: None are present to delimit the user-supplied arguments from the system instructions.
Capability inventory: The skill can execute shell commands via the paseo CLI and perform file system read operations.
Sanitization: The skill relies on instructional guidelines to "redact credentials... before sharing logs" rather than programmatic sanitization or filtering.
[COMMAND_EXECUTION]: The skill instructs the agent to run several diagnostic shell commands, including paseo --version, paseo daemon status --json, and paseo provider diagnostic <provider> --json to collect runtime information.
[DATA_EXFILTRATION]: The instructions direct the agent to read local files that may contain sensitive information, such as ~/.paseo/config.json and daemon.log. While intended for troubleshooting, these files often contain configuration details or tokens.
[EXTERNAL_DOWNLOADS]: The skill fetches documentation and indices from the vendor's official site at https://paseo.sh/llms.txt and associated markdown files to ensure current information is used.

paseo-help