new-global-command

Pass

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The workflow in SKILL.md (Step 9) instructs the agent to execute a shell command using 'uv run' where the path and arguments are partially derived from a user-provided variable ''. Without explicit sanitization or quoting instructions, this provides an opportunity for shell command injection if a user supplies a malicious skill name containing shell metacharacters (e.g., ';', '&', or backticks).
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to the lack of boundary markers and validation for user-controlled data processed by the agent.
  • Ingestion points: User-provided 'Skill name', 'Purpose', 'Scope', and 'Inputs' in the SKILL.md workflow.
  • Boundary markers: Absent. The instructions do not specify any delimiters or safety prompts to prevent the agent from interpreting instructions embedded within the user's requirements.
  • Capability inventory: The skill possesses file-writing capabilities (creating SKILL.md and scripts in 'config/codex/skills/') and shell execution capabilities ('uv run', 'make').
  • Sanitization: There is no instruction to validate, escape, or filter the user-supplied strings before they are written to disk or executed in the shell.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 24, 2026, 07:03 AM
Security Audit — agent-trust-hub — new-global-command