paw-mkt-product-context

Pass

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill includes instructions to fetch and install the agent-browser tool from a trusted repository if it is missing from the environment.
  • [COMMAND_EXECUTION]: The agent is prompted to execute shell commands to install dependencies and set up the Playwright environment for research tasks.
  • [PROMPT_INJECTION]: The skill ingests untrusted data from existing project files and live websites, creating an indirect prompt injection surface. 1. Ingestion points: Reads SOSTAC phase files from the project directory and scrapes content from public URLs. 2. Boundary markers: Uses a structured 12-section template for output but lacks explicit delimiters to separate external data from system instructions. 3. Capability inventory: Performs file-read and file-write operations (SKILL.md, references/*.md) and executes network requests via browser automation. 4. Sanitization: The skill does not define methods for sanitizing or filtering external content before processing it.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 31, 2026, 02:59 PM