The Agent Skills Directory

[PROMPT_INJECTION]: The skill is programmed to ingest data from external context files like .agents/product-marketing.md to customize its tracking advice. This pattern introduces a vulnerability surface for indirect prompt injection, where an attacker who can modify these local project files could influence the agent's behavior.
Ingestion points: The skill instructions (SKILL.md) direct the agent to read .agents/product-marketing.md, .claude/product-marketing.md, or the legacy product-marketing-context.md.
Boundary markers: There are no explicit instructions or delimiters provided to the agent to treat the content of these context files as untrusted data or to ignore embedded instructions.
Capability inventory: The skill is focused on generating implementation documentation and code snippets; it does not appear to utilize tools for system command execution, file modification, or network requests.
Sanitization: No sanitization or validation logic is specified for the content retrieved from the context files.

analytics