gh-issue
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection (Category 8).
- Ingestion points: The skill analyzes 'conversation context' and 'current conversation context' (as described in SKILL.md) to generate the content of GitHub issues.
- Boundary markers: There are no specific instructions or delimiters mentioned to prevent the agent from obeying instructions embedded within the conversation context.
- Capability inventory: The delegated agent uses the GitHub CLI ('gh'), which involves system command execution to create issues and sub-tasks.
- Sanitization: The documentation does not describe any sanitization, validation, or filtering of the conversation data before it is interpolated into commands.
- [COMMAND_EXECUTION]: The skill relies on the execution of system commands through a CLI tool.
- Evidence: The 'Agent Delegation' section states the agent uses the 'GitHub CLI (gh)' for issue creation, which is a powerful capability if the inputs are manipulated by a malicious actor.
Audit Metadata