xctf-data

Pass

Audited by Gen Agent Trust Hub on Apr 27, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches athlete profiles, meet results, and team rosters from tfrrs.org and news content from thestridereport.com as part of its core functionality.
  • [EXTERNAL_DOWNLOADS]: Downloads and installs the sports-skills Python package from the vendor's repository at github.com/machina-sports/sports-skills.git during the setup phase.
  • [COMMAND_EXECUTION]: Executes the sports-skills CLI tool to perform data retrieval and search operations.
  • [COMMAND_EXECUTION]: Includes a bash script scripts/validate_params.sh designed to ensure required arguments are provided to the sports data commands.
  • [REMOTE_CODE_EXECUTION]: Instructs the agent to perform a pip install from a remote Git URL, which involves the execution of the package's installation scripts.
  • [PROMPT_INJECTION]: Indirect prompt injection surface identified:
  • Ingestion points: External sports data and news summaries are ingested via the get_athlete_profile and get_news tools.
  • Boundary markers: No explicit markers or warnings are used to prevent the agent from interpreting instructions that might be embedded in the athlete names, meet descriptions, or news titles.
  • Capability inventory: The skill has the capability to execute arbitrary CLI commands provided by the sports-skills package.
  • Sanitization: No specific sanitization or validation of the retrieved external content is described in the provided files.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 27, 2026, 03:27 PM