The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill demonstrates patterns for ingesting external data into agent-driven network calls. 1. Ingestion points: userId, email, and input variables in references/frontend-integration.md. 2. Boundary markers: Absent; data is interpolated directly into strings. 3. Capability inventory: The skill uses sdk.client.fetch in SKILL.md and references/frontend-integration.md to perform network operations. 4. Sanitization: None shown in the provided examples. This represents a standard surface for indirect prompt injection where malicious data in a user profile or API response could attempt to influence agent behavior.
[Prompt Injection] (SAFE): No malicious instructions, bypass attempts, or behavioral overrides were detected in the skill logic.
[Credential Safety] (SAFE): No hardcoded API keys, tokens, or sensitive file paths (e.g., .env, .ssh) were found. Placeholders like userId are used safely.
[Remote Code Execution] (SAFE): No patterns of downloading and executing remote scripts (e.g., curl|bash) were detected. The skill relies on standard, well-known libraries.

building-storefronts