The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process historical data from databases, which constitutes untrusted external input. This creates an attack surface where malicious instructions embedded in database records could potentially influence agent behavior during processing or restoration operations.\n
Ingestion points: Database tables accessed via psql and mysql CLI tools as described in SKILL.md.\n
Boundary markers: The instructions do not define specific delimiters or security prompts to prevent the agent from interpreting record content as instructions.\n
Capability inventory: The skill utilizes powerful capabilities including shell access for database and cloud providers (Bash) and file system operations (Write, Edit).\n
Sanitization: No explicit sanitization or validation of database record content is mentioned in the archival process.\n- [PERSISTENCE]: The skill explicitly instructs the implementation of automated archival routines using system cron jobs or database schedulers. While this is the intended functionality for data retention, it represents the creation of persistence mechanisms on the host system.\n- [COMMAND_EXECUTION]: The skill requires the use of multiple shell-based tools (psql, mysql, aws, az, gsutil, gzip) to perform its core functions of data identification, extraction, compression, and cloud migration.

archiving-databases