The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external, potentially untrusted data from Notion to generate positioning hypotheses, creating a surface for indirect prompt injection.
Ingestion points: The agent reads data from Notion pages, including client verbatim quotes, competitor analysis, and market insights (referenced as steps 01, 02, 03, and 05).
Boundary markers: The skill includes a mandatory checkpoint, 'GATE 2.5', where the agent must present its hypotheses to the user for confirmation or replacement before proceeding to subsequent tasks.
Capability inventory: The skill is designed to read from and write to Notion. It does not contain capabilities for arbitrary command execution, file system modification, or network requests to unknown domains.
Sanitization: The instructions emphasize a structured '8-step spine' and stage-gating to ensure output is grounded in specific business rules rather than raw external input alone.
[SAFE]: The skill follows professional business automation practices. It utilizes internal reference documents (agency-stages.md and gtm-frameworks.md) attributed to the skill's author to guide its logic. No obfuscation, hardcoded credentials, or persistence mechanisms were detected. The use of the Notion platform is consistent with the skill's primary purpose.

gtm-positioning