opencode-permission

Pass

Audited by Gen Agent Trust Hub on May 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a bundled Python script, manage_permission.py, to modify local configuration files. This is the intended purpose of the tool and is performed without elevated system privileges.
  • [PROMPT_INJECTION]: The skill provides a mechanism for an agent to modify its own security boundaries, which constitutes an indirect prompt injection surface.
  • Ingestion points: User-supplied command strings and permission actions are processed from the prompt and written to the configuration file.
  • Boundary markers: No specific delimiters are used to isolate user-provided command patterns from the agent's internal configuration logic.
  • Capability inventory: The script has the capability to write to ~/.config/opencode/opencode.jsonc, which governs the agent's permission policy for all other tool calls.
  • Sanitization: The script validates that the action parameter matches expected values (allow, ask, deny) but does not sanitize the content of the command patterns themselves.
Audit Metadata
Risk Level
SAFE
Analyzed
May 6, 2026, 01:01 AM
Security Audit — agent-trust-hub — opencode-permission