The Agent Skills Directory

[PROMPT_INJECTION]: The skill analyzes user codebases to generate infrastructure recommendations, which introduces a surface for indirect prompt injection. Malicious instructions or metadata within the project files could attempt to influence architectural decisions or weaken security configurations.
Ingestion points: Codebase analysis described in SKILL.md (Step 1) and references/defaults.md.
Boundary markers: Absent; there are no specific instructions for the agent to treat findings from the codebase scan as untrusted data or to ignore embedded natural language commands.
Capability inventory: Possesses capabilities for command execution and infrastructure-as-code deployment via shell tools.
Sanitization: Absent; the skill does not specify validation or sanitization routines for data extracted during the codebase scan.
[REMOTE_CODE_EXECUTION]: Generates and executes Infrastructure-as-Code (IaC) templates. This functionality is protected by a requirement for explicit user confirmation before any deployment command is run.
[COMMAND_EXECUTION]: Executes shell commands for codebase analysis and security scanning as part of the deployment workflow.
[EXTERNAL_DOWNLOADS]: Recommends the use of reputable security tools like checkov, cfn-nag, and truffleHog to audit infrastructure and prevent credential leakage. These references are to well-known industry services.

deploy