four-color-cards
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests markdown content and user intent to generate analytical output via LLMs, which is a standard surface for indirect prompt injection. This is evaluated as safe as it represents the core functionality of the skill without malicious intent.
- Ingestion points: The
markdownandintentfields defined inSKILL.mdserve as ingestion points for untrusted data. - Boundary markers: The skill instructions do not define specific delimiters or warnings to isolate external data from internal instructions.
- Capability inventory: The skill executes Python code via
scripts/run_four_color_cards.pywhich performs file system writes to a specific output directory and interacts with LLM agents. - Sanitization: Input data is processed directly without explicit sanitization or filtering logic.
Audit Metadata