The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local binary at .slipbox/bin/slipbox using shell interpolation for arguments like <draft-slug> and <tension description>. These values are derived from external source content and AI-generated summaries. The lack of sanitization or escaping for these variables before they are passed to the shell creates a potential vector for command injection.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from external sources and uses it to drive subsequent tool calls and shell executions.
Ingestion points: Content from user-named resources, clipped resources, and existing markdown files in the /literature/ directory.
Boundary markers: Absent. The skill does not employ delimiters or specific instructions to isolate source content from the agent's logic.
Capability inventory: The skill has the capability to write to the filesystem and execute the local slipbox binary.
Sanitization: Absent. Extracted data is interpolated directly into command arguments and markdown files without validation or filtering.

ground-claim