scientific-critical-thinking
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of instructional Markdown and reference documentation. No scripts, binaries, or automated tools are included. The suggested use of
grepfor searching references inSKILL.mdis a standard local operation within the skill environment.\n- [NO_CODE]: There are no executable files, scripts (Python, JavaScript, etc.), or configuration files for automated tools included in the skill package.\n- [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for external data (research papers and claims) mentioned inSKILL.md.\n - Ingestion points: User-provided research papers, preprints, and scientific assertions for analysis.\n
- Boundary markers: None explicitly defined in the instructions to separate paper content from agent instructions.\n
- Capability inventory: None. The skill does not define any scripts or tools that perform network operations or file writes.\n
- Sanitization: None specified for input data.
Audit Metadata