The Agent Skills Directory

[SAFE]: The skill provides a structured methodology for marketing strategy without including any executable code or malicious instructions.
[SAFE]: Data handling practices are secure; the skill instructs users to store API keys in .env files rather than hardcoding them.
[SAFE]: External operations, such as publishing a final plan to a GitHub repository, are presented as optional user-initiated actions and involve standard project management tools.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from client materials and third-party APIs (Ahrefs, Stripe, etc.) to inform the generated plan. While this presents a theoretical attack surface, the risk is inherent to the skill's purpose and is mitigated by the structured planning process and the absence of automated command execution based on that data.
Ingestion points: Reads user-provided files in ~/marketing-plans/{client-slug}/materials/ and pulls data from external APIs (Ahrefs, Google Analytics, Stripe, Customer.io, etc.).
Boundary markers: The skill uses a structured multi-phase workflow (INIT, REVIEW, FINALIZE) which acts as a logical boundary.
Capability inventory: Capable of writing files to the local file system (final_plan.md) and pushing content to a user-specified GitHub repository.
Sanitization: No explicit sanitization of input data is described, though the output is structured markdown intended for human review.

marketing-plan