The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from external specifications and codebases to generate project tickets. * Ingestion points: The skill reads "whatever is already in the conversation", fetched reference specs, maps, and project codebase files (Steps 1 and 2). * Boundary markers: No explicit delimiters or instructions are used to distinguish between system instructions and processed data. * Capability inventory: The skill performs file reads to gather context and file writes to create ticket markdown files in the ".scratch/" directory. * Sanitization: There is no evidence of sanitization or filtering of the content being processed and written to the ticket files.

to-tickets