The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional Markdown, project templates, and evaluation criteria. Analysis of the provided files (SKILL.md, content-architecture.md, templates.md, evals.json) confirmed that no executable scripts, binary files, or network-active commands are included.- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external project data. Evidence chain: 1. Ingestion points: Reads from .agents/marketing-strategy.md, .agents/product-marketing.md, and .agents/marketing-learnings.md (SKILL.md). 2. Boundary markers: Explicit instruction to "Refuse to misrepresent a competitor's features or fabricate a switcher testimonial, even if asked" (SKILL.md). 3. Capability inventory: Generates text content and structured YAML profiles. 4. Sanitization: None identified. The risk is mitigated by the skill's focus on generated content rather than executable output.- [DATA_EXFILTRATION]: The skill accesses project-specific strategy and marketing files to inform its generation process. These operations are local to the environment and are necessary for the skill's core functionality. No patterns of data transmission to external or untrusted domains were detected.

competitors