marketing-context

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's "Auto-Draft" feature scans the codebase, making it susceptible to malicious instructions hidden in files like README.md or package.json that could influence the resulting marketing context.
  • Ingestion points: Codebase files including README, landing pages, package.json, and documentation as specified in SKILL.md.
  • Boundary markers: Absent; there are no specific instructions to delimit or ignore embedded prompts in the source files.
  • Capability inventory: Local script execution (scripts/context_validator.py) and file system write access to the .claude/ directory.
  • Sanitization: Absent; the skill drafts content directly from repo content without explicit filtering.
  • [COMMAND_EXECUTION]: The skill executes a local Python utility (scripts/context_validator.py) to verify the completion score of the marketing document.
  • [DATA_EXPOSURE]: The skill requires read access to repository configuration and documentation to extract product overview and target audience information.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 05:51 PM
Security Audit — agent-trust-hub — marketing-context