The Agent Skills Directory

[DATA_EXFILTRATION]: The skill logic in scripts/main.cjs resolves file paths for reading content and writing reports using path.resolve(PROJECT_ROOT, path). This implementation fails to verify that the resulting paths are contained within the project directory, potentially allowing for directory traversal (e.g., using ../../ in the contentPath or outputPath parameters) to access or overwrite files outside the intended project scope.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted data (marketing copy, design assets) which is then analyzed by the AI without sufficient protective measures.
Ingestion points: Content enters the skill via the contentPath or contentText properties in scripts/main.cjs and is processed according to instructions in SKILL.md (Steps 1, 2, and 5).
Boundary markers: The agent's instructions do not utilize delimiters (e.g., XML tags) or specific directives to ignore embedded instructions within the audited content.
Capability inventory: The agent is equipped with high-privilege file system tools including Read, Write, Edit, Grep, and Glob.
Sanitization: There is no sanitization, escaping, or validation of the external content before it is interpolated into the agent's context for tone and visual analysis.

brand-compliance