The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its context-loading mechanism.
Ingestion points: The skill instructions explicitly tell the agent to read context from external files such as .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md before proceeding (documented in SKILL.md under the 'Initial Assessment' section).
Boundary markers: There are no instructions to the agent to treat the content of these files as untrusted or to ignore any instructions found within them.
Capability inventory: The agent uses the data from these files to generate email sequences, marketing strategies, and copy, which could be manipulated by an attacker who has write access to the project's repository/file system.
Sanitization: No sanitization, escaping, or validation of the content from the external files is performed before the agent processes it.

emails