The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to read external context files such as .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md to inform its recommendations. If these files are sourced from an untrusted environment or modified by an attacker, they could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The skill instructions in SKILL.md explicitly direct the agent to read product marketing context files from the local filesystem.
Boundary markers: There are no explicit instructions or delimiters defined to isolate the content of these context files from the agent's primary instructions.
Capability inventory: The skill provides optimization recommendations and design guidance based on the ingested data.
Sanitization: No validation or sanitization mechanisms are described for the content of the marketing context files.

paywalls