The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is configured to ingest data from local project files, which creates a potential surface for indirect prompt injection if these files contain malicious instructions.
Ingestion points: The SKILL.md file instructs the agent to read context from files such as .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md before proceeding with tasks.
Boundary markers: There are no explicit instructions or delimiters defined in the skill to help the agent distinguish between valid marketing data and potential malicious instructions within those ingested files.
Capability inventory: The skill operates within an environment that typically allows for file system operations and command execution, which could be targeted if an injection occurs.
Sanitization: No validation or sanitization steps are defined for the content read from the project context files.

copywriting