The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides instructions for querying raw 'signals' which are natural-language descriptions of product events. Since these signals can be triggered by external activity (e.g., error logs, web analytics), an attacker could potentially influence the content of these signals to include malicious instructions.
Ingestion points: The content column of the document_embeddings table is accessed through the posthog:execute-sql tool as described in SKILL.md.
Boundary markers: None present in the provided SQL templates to distinguish between data and instructions.
Capability inventory: The skill uses posthog:execute-sql for database access and embedText() for semantic searches.
Sanitization: No explicit sanitization or instructions to ignore instructions within the signal content are provided.

signals