copywriting

Pass

Audited by Gen Agent Trust Hub on Apr 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill operates as intended for copywriting tasks without requesting external network access or sensitive credentials.
  • [COMMAND_EXECUTION]: Includes a Python script scripts/headline_scorer.py for evaluating headlines. Analysis shows the script is self-contained, uses standard Python libraries, and does not perform any dangerous system operations or network requests.
  • [PROMPT_INJECTION]: The skill is designed to ingest data from .claude/product-marketing-context.md and user-provided descriptions, creating a surface for indirect prompt injection.
  • Ingestion points: .claude/product-marketing-context.md (read via file system) and direct user input for audience and product details.
  • Boundary markers: None; the instructions do not specify delimiters for the external context file content.
  • Capability inventory: Generation of marketing copy and execution of the local scoring script.
  • Sanitization: No validation or filtering is performed on the content of the marketing context file before it is used by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 12, 2026, 08:17 AM
Security Audit — agent-trust-hub — copywriting