The Agent Skills Directory

[SAFE]: The skill provides comprehensive guidance on remediating security alerts (W001, W011, W012) by restructuring skill content. These practices—such as version pinning, moving installs to frontmatter, and removing explicit tool names from prose—are defensive measures aligned with security best practices for AI agents.
[PROMPT_INJECTION]: The skill's primary function involves analyzing and modifying third-party skill files, which establishes an indirect prompt injection surface. A malicious skill file could attempt to influence the agent's behavior during the compliance review process.
Ingestion points: The skill processes all files within a target skill directory, including SKILL.md, references/, and assets/.
Boundary markers: No explicit boundary markers or 'ignore' instructions are provided for the content being analyzed.
Capability inventory: The skill is granted significant permissions, including Read, Edit, Write, and Bash (scoped to git, uv, and uvx).
Sanitization: No specific input validation or sanitization routines are mentioned for the ingested skill content.

snyk-agent-scan-compliance