transparent-proxy-maintenance

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect prompt injection surface (Category 8). The skill modifies Python plugin files and JSON configuration snapshots based on user-provided values for pricing, rate limits, and API structures.
  • Ingestion points: User requests specifying changes to pricing parameters, rate limits, or new API definitions in the plugin files mentioned in SKILL.md (e.g., apis/*.py).
  • Boundary markers: Absent. There are no explicit instructions to treat user-provided configuration values as untrusted or to use delimiters to isolate user input from the rest of the code.
  • Capability inventory: Use of edit_file to modify source code and configuration, and fly deploy to deploy modified code to production or test environments.
  • Sanitization: The skill prescribes functional validation and smoke testing post-deployment but lacks input sanitization for the strings being written to files.
  • [COMMAND_EXECUTION]: The skill uses administrative CLI tools to manage and verify remote infrastructure.
  • Evidence: Instructions for using fly deploy and fly ssh console to update the service and perform non-interactive verification on remote machines.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 01:04 AM