The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local shell script audit_redis_configs.sh and python modules. * Evidence: Referenced in SKILL.md workflow and implemented in audit_redis_configs.sh. * Context: Depends on a pre-configured virtual environment located at /mnt/venv_ext4/venv_render_signal_server.
[DATA_EXFILTRATION]: The skill is designed to read, display, and modify sensitive configuration data, including magic_link_tokens. * Evidence: SKILL.md specifies magic_link_tokens as a configuration object to be inspected and reconciled using MCP Redis tools. * Evidence: audit_redis_configs.sh retrieves and prints these tokens for comparison.
[PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by processing external data from Redis and local files. * Ingestion points: Redis keys fetched via MCP get and debug/*.json fallback files. * Boundary markers: None present; the skill assumes the integrity of the data it reconciles. * Capability inventory: Redis set and delete tools, shell execution, and python script execution. * Sanitization: No validation or sanitization of the JSON payloads is mentioned before processing.

redis-config-guardian