The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data provided by the user (such as stack traces, error reports, and pasted logs) and external sources (databases, Redis, and configuration files). This data is processed in a context where the agent has access to highly capable tools like Bash and Edit.\n
Ingestion points: User-supplied messages containing symptoms and reproduction steps, and system data read via the Read tool or database queries.\n
Boundary markers: The skill does not explicitly define delimiters or specific instructions to ignore embedded directives in the data being investigated.\n
Capability inventory: The agent is authorized to use Bash for shell command execution and Edit for modifications across the project files.\n
Sanitization: The instructions include a specific security requirement to strip hostnames, IPs, and customer data before performing external searches via WebSearch, which mitigates data leakage risks.\n- [COMMAND_EXECUTION]: The skill uses the Bash tool to perform complex diagnostic tasks including git bisect for regression identification and environment variable inspection (env | grep APP_). These commands are standard for debugging but involve broad read and execute access to the system environment.

investigate