The Agent Skills Directory

[PROMPT_INJECTION]: The skill defines a workflow that ingests untrusted external data, creating a surface for indirect prompt injection.
Ingestion points: Data enters the agent context through the WebSearch tool (used for market analysis) and the Read/Grep tools (used for processing user feedback and project documentation).
Boundary markers: The skill instructions do not provide explicit delimiters or "ignore embedded instructions" warnings for handling data retrieved from these external or user-provided sources.
Capability inventory: The skill has access to powerful capabilities including Bash, Task, Write, and Edit for file modification and command execution.
Sanitization: There is no evidence of sanitization, validation, or escaping logic for external inputs before they are interpolated into prompts or used in workflows.
[COMMAND_EXECUTION]: The skill is configured with high-privilege execution tools to facilitate its core functions.
Evidence: The allowed-tools configuration explicitly enables Bash and Task. These tools are intended for project initialization, monorepo management, and executing infrastructure-as-code (Terraform/Kubernetes) commands as described in the tech stack and project structure sections.

enterprise