aeon-vuln-scanner
Warn
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill performs manual triage by reading 30-50 lines of code context from external, untrusted GitHub repositories. A malicious repository could include comments containing instructions designed to mislead the agent into ignoring vulnerabilities or performing unintended actions.
- Ingestion points: code context read during triage (SKILL.md).
- Boundary markers: Absent; the agent reads raw code blocks without protective delimiters or instructions to ignore embedded commands.
- Capability inventory: GitHub API access via
gh api(POST requests), shell command execution for scanners, and repository writing for PR creation. - Sanitization: None specified; the agent directly processes the retrieved code lines.
- [COMMAND_EXECUTION]: The skill uses the
$REPOvariable within a shell command (gh api -X POST "/repos/$REPO/security-advisories"). Since the repository name originates from external sources like the GitHub trending list, a maliciously crafted repository name could potentially trigger command injection if the shell environment does not adequately sanitize the variable.
Audit Metadata