paper-glossary

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a robust cryptographic provenance chain that validates data integrity across the proposal, review, triage, and writing phases. Every step re-computes SHA-256 digests to ensure that the final glossary entries match the user-selected candidates and the original source paper material, providing a strong defense against indirect prompt injection.
  • [SAFE]: All file operations are constrained to valid Obsidian vaults by verifying the existence of the .obsidian directory in the target path hierarchy.
  • [SAFE]: The skill operates entirely on local infrastructure, utilizing configuration stored in ~/.paper-glossary/config.json without any detected network activity or external data exfiltration.
  • [SAFE]: File writing procedures follow security best practices by using atomic replacements (writing to a temporary file before using os.replace) and preserving existing file permissions.
  • [SAFE]: The integration test suite uses subprocess.run to invoke the skill's own scripts, which is a standard testing practice and does not involve any unsanitized or remote input.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 10:29 AM