The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface by processing external research data and a local configuration file (.claude/product-marketing-context.md). 1. Ingestion points: The agent is instructed to read local project context and mine external review sites (G2, Capterra). 2. Boundary markers: The skill instructions emphasize honesty and verification, which serve as natural constraints on generated content. 3. Capability inventory: The skill is restricted to generating markdown text; it contains no subprocess calls, network exfiltration capabilities, or dynamic execution tools. 4. Sanitization: No explicit sanitization logic is provided, but the risk is mitigated by the lack of powerful execution tools.

competitor-alternatives