aibijia-price-comparison
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill scrapes price and product data from various external AI token reseller platforms which are untrusted third-party sources. This creates a surface for indirect prompt injection if the scraped strings are later interpreted by an agent as instructions. 1. Ingestion points: KawangScraper.fetch_page in SKILL.md fetches data from external reseller URLs. 2. Boundary markers: No explicit delimiters or instructions are used to separate scraped content from agent prompts. 3. Capability inventory: The skill performs file writing (save_price_snapshot), network POST requests (submit_vendor), and automated Telegram messaging (send_price_alert). 4. Sanitization: The logic uses basic HTML parsing but lacks validation to strip potential prompt injection payloads from vendor-controlled fields.
- [COMMAND_EXECUTION]: The documentation provides setup instructions to execute shell commands from a non-verified source. Evidence: Instructions include git clone https://github.com/ka-pi-ba-la/AIbijia.git followed by dependency installation via pip and npm.
- [EXTERNAL_DOWNLOADS]: The skill is designed to perform network requests to scrape data from third-party card-selling platforms and communicates with aibijia.org for community vendor submissions.
Audit Metadata