paper-write

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted research notes (PAPER_PLAN.md, NARRATIVE_REPORT.md) which are interpolated into prompts. 1. Ingestion points: User-provided markdown files in the local workspace. 2. Boundary markers: Absent; instructions do not enclose ingested content in delimiters or include 'ignore instructions' warnings. 3. Capability inventory: Bash, Write, Edit, Agent. 4. Sanitization: Absent; content is processed directly. This creates a surface for indirect prompt injection where malicious instructions in research notes could influence agent behavior.
  • [EXTERNAL_DOWNLOADS]: The skill fetches bibliography metadata from dblp.org and doi.org. These are well-known academic services used to prevent citation hallucinations and ensure reference accuracy.
  • [COMMAND_EXECUTION]: Executes shell commands and an embedded Python script to organize the LaTeX project structure, manage backups, and perform citation hygiene checks within the local environment.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 06:06 PM
Security Audit — agent-trust-hub — paper-write