natural-center-extraction
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides instructional logic for rhetorical analysis and does not contain any executable scripts, remote downloads, or obfuscated content. Its behavior is consistent with its stated goal.
- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it is designed to ingest and process untrusted external artifacts.
- Ingestion points: Source artifacts such as essays, transcripts, and session logs as defined in the
inputssection ofSKILL.md. - Boundary markers: Absent; there are no instructions to the agent to disregard instructions potentially embedded within the source artifacts.
- Capability inventory: The skill utilizes filesystem access (
reads-filesystemandcreates-files) to perform content analysis and generate output. - Sanitization: None; the skill is instructed to return verbatim segments of the source material.
Audit Metadata