research-summarizer

Pass

Audited by Gen Agent Trust Hub on Apr 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill documentation instructs users to execute local Python scripts (scripts/extract_citations.py, scripts/format_summary.py) and a shell script (scripts/convert.sh) as part of its operational workflow.
  • [EXTERNAL_DOWNLOADS]: Installation instructions recommend cloning a repository from GitHub (github.com/alirezarezvani/claude-skills.git) and using an external package manager (clawhub) to install the skill.
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its primary function of processing and summarizing external, untrusted research documents.
  • Ingestion points: Untrusted data is ingested via scripts/extract_citations.py (reading from files or stdin) and directly by the agent when prompted to summarize user-provided documents.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are included in the prompt templates or scripts.
  • Capability inventory: The skill involves local command execution (Python) and file read access.
  • Sanitization: The extraction script uses regular expressions for pattern recognition but does not perform sanitization or filtering of the source text content.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 12, 2026, 08:17 AM
Security Audit — agent-trust-hub — research-summarizer