The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted PDF documents from local storage and S3 buckets, creating a surface for indirect prompt injection.
Ingestion points: Reads and extracts text from PDF files via the asta pdf-extraction command (SKILL.md).
Boundary markers: The instructions do not include explicit delimiters or warnings to the agent to disregard instructions embedded within the extracted document content.
Capability inventory: The skill uses Bash for command execution and filesystem access for reading and writing (SKILL.md).
Sanitization: There is no mention of sanitization or filtering of the extracted text before it is returned to the agent context.
[COMMAND_EXECUTION]: The skill relies on shell commands for its primary operations and workspace management.
Evidence: The instructions include the use of asta pdf-extraction, find, mktemp, mv, cp, and rm via the Bash tool (SKILL.md).
Risk: These operations involve interpolating paths and variables into shell commands, which can be vulnerable to injection if inputs are maliciously crafted.

pdf-extraction