linkfox-temu-price-global
Warn
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: MEDIUMPERSISTENCECREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PERSISTENCE]: The skill provides instructions in
references/onboarding.mdfor users to manually add environment variable exports to shell configuration files (e.g.,~/.zshrc,~/.bashrc). This is a persistence mechanism intended to maintain access to LinkFox API keys across system sessions.\n- [CREDENTIALS_UNSAFE]: The skill manages sensitive authentication data by reading from and writing to a local token store located at~/.linkfox/temu-access-tokens.json. Accessing and storing credentials in hidden system directories is a sensitive operation performed for the skill's core authorization functionality.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external Temu and LinkFox API responses and writes the full JSON output to the local workspace for agent processing, creating a vulnerability surface for indirect prompt injection.\n - Ingestion points: Data received from the LinkFox gateway and Temu platform APIs via the
temu/proxyendpoint.\n - Boundary markers: No explicit delimiters or safety warnings are implemented to isolate untrusted API content in the generated JSON files.\n
- Capability inventory: The skill includes file system write access (
_temu_common.py) and network egress capabilities (urllib.requestandrequests.post).\n - Sanitization: Content from external APIs is written to disk without sanitization or validation of its contents.\n- [COMMAND_EXECUTION]: Multiple Python scripts (e.g.,
temu_proxy.py,global_price_priceorder_query.py) execute network requests and perform local file operations based on JSON parameters provided through the command line.
Audit Metadata