os-environment-probe

Pass

Audited by Gen Agent Trust Hub on Aug 9, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill executes external command-line interfaces including 'gh', 'agy', and 'cursor' via the 'Bash' tool to verify their availability using non-interactive flags. It also references a support script 'scripts/probe_environments.sh' which is intended for local execution but is not included in the skill's file list.
  • [INDIRECT_PROMPT_INJECTION]: The skill populates a markdown configuration file 'environment.md' that is consumed by downstream skills, creating a data dependency for future agent actions. 1. Ingestion points: 'context/memory/environment.md' is read by 'os-evolution-planner' and 'os-architect'. 2. Boundary markers: The file uses structured tables but lacks explicit sanitization instructions for the 'Other' user input category. 3. Capability inventory: Downstream skills utilize the file contents to construct command-line arguments for 'run_agent.py'. 4. Sanitization: No explicit validation or escaping of the environment names is performed before writing to the profile.
  • [DYNAMIC_EXECUTION]: The skill facilitates dynamic tool selection by generating configuration data that dictates which execution backends are used by the agent during subsequent tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 9, 2026, 12:57 PM
Security Audit — agent-trust-hub — os-environment-probe