claude-marketing-skills
Warn
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONPERSISTENCE
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructs users to clone a repository and run multiple setup scripts (e.g.,
python skill-packs/scripts/setup-paid-media.py). This practice involves downloading and executing unverified code from an external source.\n- [EXTERNAL_DOWNLOADS]: Skills and dependencies are retrieved from a non-whitelisted GitHub repository (thatrebeccarae/claude-marketing). The skill also utilizesnpx -yto fetch and execute packages directly from the NPM registry.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources that could contain malicious instructions.\n - Ingestion points: Tools such as
BrandDNAExtractorread content from arbitrary URLs, and theCompetitorAnalystscrapes public ad libraries.\n - Boundary markers: Documentation does not mention the use of delimiters or specific instructions to ignore embedded prompts in the ingested data.\n
- Capability inventory: The skill has the ability to write files, perform network operations to various marketing APIs, and execute shell commands.\n
- Sanitization: No evidence of sanitization or filtering is provided for the content retrieved from external URLs.\n- [COMMAND_EXECUTION]: The installation process involves shell commands to copy directories (
cp -r) into the agent's internal configuration folders and runningnpxcommands that interact with the system.\n- [PERSISTENCE]: Instructions involve copying files into~/.claude/skills/and modifying~/.claude/mcp_settings.json, which ensures that external code and configurations persist across agent sessions.
Audit Metadata