The Agent Skills Directory

[PROMPT_INJECTION]: The skill includes an indirect prompt injection surface as it is designed to ingest and process untrusted documents (decks, transcripts, audits) provided by external clients.
Ingestion points: The agent is directed to read all materials stored in the ~/marketing-plans/{client-slug}/materials/ directory.
Boundary markers: The skill instructions do not explicitly mandate the use of delimiters or 'ignore' instructions for content found within these external materials.
Capability inventory: The agent has the capability to write to the local file system and interact with high-privilege third-party APIs via MCP tools (e.g., Stripe, GitHub).
Sanitization: No explicit content sanitization or instruction-filtering is defined for the ingestion phase.
[DATA_EXFILTRATION]: The skill is designed to access and analyze sensitive business data from services like Stripe and Google Analytics.
Trusted Sources: All external service integrations target well-known and trusted technology providers.
Secure Configuration: The documentation correctly advises storing API credentials in .env files, which is an industry-standard safe practice.
Leak Prevention: The workflow includes a mandatory 'verification pass' during the finalization phase to identify and remove machine-specific paths or accidental credential exposure from the output plan.
[SAFE]: No malicious obfuscation, base64-encoded payloads, unauthorized command execution, or persistence mechanisms were detected. The skill's behavior and requested tool access are strictly aligned with its primary purpose of marketing strategy and research.

marketing-plan