The Agent Skills Directory

[SAFE]: No security issues were identified during analysis. The skill defines a constrained workflow for generating founder-focused LinkedIn content. It does not contain any instructions for network exfiltration, remote code execution, or privilege escalation. All referenced tools and external links are consistent with the skill's stated purpose of content generation and professional branding.
[INDIRECT_PROMPT_INJECTION]: The skill includes an attack surface for indirect prompt injection as it ingests user-provided topics and 'receipts' to generate posts.
Ingestion points: Found in the 'Activation' and 'Capture the specifics' sections of SKILL.md where the agent accepts user input.
Boundary markers: The workflow enforces a rigid output structure (Hook, Body, Receipt block, Close) and a voice self-check rubric, which provides a level of control over the generated content.
Capability inventory: The skill has access to Read, Write, and Grep tools; however, it lacks network access or shell execution capabilities that would be required to leverage an injection for high-severity impact.
Sanitization: The skill utilizes a 'Banned patterns' list to explicitly refuse the generation of low-quality or suspicious phrases, providing a basic level of input/output filtering.

founder-content