The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute directory creation commands (mkdir -p). The directory names are derived from the project slug, which is generated from the 'brand-name' found in the 'portfolio.json' file or provided by the user. While the instructions specify converting the name to 'kebab-case', this represents a point where external data influences command execution.- [EXTERNAL_DOWNLOADS]: The skill is configured with access to network-enabled tools (WebSearch, WebFetch). While the primary workflow focuses on local project discovery and configuration, these tools provide the capability to retrieve external content.- [PROMPT_INJECTION]: The skill processes untrusted data from local JSON files (portfolio.json, tips-project.json, and market definitions) to configure the marketing project. This creates an indirect prompt injection surface where malicious content inside these files could influence agent behavior during the setup process. Ingestion points: 'SKILL.md' Step 0 and Step 2. Boundary markers: Not explicitly defined for file content. Capability inventory: Includes 'Bash', 'Write', 'Edit', and 'WebFetch'. Sanitization: The instructions mandate 'kebab-case' formatting for the project slug used in shell commands.

marketing-setup