The Agent Skills Directory

[SAFE]: The skill implements a robust design workflow that emphasizes user confirmation and brand fidelity. It uses project-specific files as intended for creative tasks.
[COMMAND_EXECUTION]: Utilizes a restricted Bash command scoped to mktg status for verifying project state, which does not present a privilege escalation risk.
[PROMPT_INJECTION]: The skill processes data from brand and marketing files (brand/*.md, marketing/content-specs/*.yaml) to inform its creative decisions. This creates a surface for indirect prompt injection if these files are compromised by an attacker.
Ingestion points: Reads project brand files and content specifications.
Boundary markers: Content is interpolated into subagent prompts without explicit delimiters or 'ignore' instructions.
Capability inventory: Spawns subagents with specific roles and performs file system writes.
Sanitization: Lacks explicit content validation or sanitization for external data.

paper-marketing