The Agent Skills Directory

[DATA_EXFILTRATION]: The skill documentation identifies that credentials and configuration are stored at ~/.config/mlwcli/auth.toml. This is the standard configuration path for the mlwcli tool's persistent state.
[PROMPT_INJECTION]: The skill processes data from external Miniflux, Linkding, and Wallabag services, which constitutes an indirect prompt injection surface.
Ingestion points: External data enters the agent context via the link list, feed list, entry list, and page list commands.
Boundary markers: The skill encourages the use of the --jq and --json flags, which provide structured data boundaries that help differentiate content from instructions.
Capability inventory: The skill has the capability to execute the mlwcli binary and read/write to its own configuration files.
Sanitization: The use of structured output (JSON/JQ) ensures that the agent receives data in a parsed format, though the semantic content of titles and notes remains unvalidated.

mlwcli