Skills
Navigate
skills/rudra-ravi/frida-skills/frida-codeshare-search/Gen Agent Trust Hub

frida-codeshare-search

Fail

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches search results and script project pages from the Frida CodeShare community repository (codeshare.frida.re).
  • [REMOTE_CODE_EXECUTION]: Loads and executes community-contributed JavaScript instrumentation scripts using the frida command-line tool, including the use of the --codeshare flag for direct remote code execution.
  • [COMMAND_EXECUTION]: Utilizes curl for network requests, python3 for parsing downloaded HTML via the codeshare_extract.py script, and frida for application instrumentation.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting and executing scripts from external third-party sources.
  • Ingestion points: HTML files (codeshare-search.html, codeshare-project.html) downloaded from codeshare.frida.re.
  • Boundary markers: The skill instructions include explicit rules to review source code using sed and confirm target authorization before execution.
  • Capability inventory: Networking (curl), file system writes (codeshare_extract.py), and dynamic instrumentation (frida).
  • Sanitization: Relies on manual inspection by the agent rather than automated sanitization of the downloaded script content.
Recommendations
  • HIGH: Downloads and executes remote code from: https://codeshare.frida.re/search/, https://codeshare.frida.re/@Eltion/instagram-ssl-pinning-bypass/ - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 24, 2026, 11:23 AM
Security Audit — agent-trust-hub — frida-codeshare-search