competitive-intel-agent
Warn
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill instructions direct the agent to construct and execute shell commands for generating images and PDF reports. These commands incorporate user-controlled data such as product names and competitor names directly into CLI arguments, which could lead to command injection if malicious strings are supplied.
- Evidence: SKILL.md contains command templates in Step 4 and Step 6 that interpolate [COMPETITORS] and [Your Product] into python3 command strings.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from the web (competitor websites, reviews, and sentiment) and user inputs, which are then used as context for analysis agents. The workflow lacks clear boundary markers or sanitization logic to handle potentially malicious content within these external sources.
- Ingestion points: Competitor URLs and names provided by users, and web-based research conducted by sub-agents like the market-position-analyst.
- Boundary markers: No explicit delimiters or instructions are provided to the agents to ignore embedded instructions in the research data.
- Capability inventory: The agent can execute shell commands via python3 scripts and write files (PNG and PDF) to the local filesystem.
- Sanitization: No explicit logic is defined to validate or escape external content before it is processed or interpolated into command-line prompts.
Audit Metadata