paw-mkt-guerrilla
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for the environment to install the
agent-browsertool from the Vercel Labs GitHub organization and official package registries. These are recognized trusted sources and are used to facilitate marketing research tasks. - [COMMAND_EXECUTION]: The instructions include standard shell commands for checking tool availability (e.g.,
node,git,python3) and performing browser-based research. These operations are restricted to the local environment and intended for legitimate marketing auditing and data gathering. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an "Existing Local Codebase" and "Live Website URL Audit" context router, which allows the agent to ingest data from external sources and project files. While this creates a potential surface for indirect prompt injection, the skill includes explicit instructions for human oversight, rigorous risk assessment, and validation of strategic alignment before any tactics are executed.
- [SAFE]: The skill implements a robust governance framework, including a mandatory "Risk Assessment Gate" and "Legal & Ethical Boundaries" section. These instructions explicitly prohibit illegal activities, trademark infringement, and deceptive practices (astroturfing), while requiring documented risk/reward scoring (ICE-R) for all proposed tactics.
Audit Metadata