paper-memory-builder
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions suggest running a local Python script
scripts/check_claims_schema.pyto validate generated YAML outputs against a JSON schema. This is a standard developer workflow for data integrity. - [PROMPT_INJECTION]: The skill ingests untrusted manuscript drafts (Word, LaTeX, Markdown) which serves as an attack surface for indirect prompt injection. The skill mitigates this risk by requiring the agent to copy claim text verbatim and forbidding rephrasing or fabrication.
- [EXTERNAL_DOWNLOADS]: The skill references external JSON schema definitions from
json-schema.organd the author's own GitHub repository for metadata and validation purposes. No executable code is downloaded or executed from these remote sources at runtime.
Audit Metadata