agent-swarm-issue
Fail
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: CRITICALNO_CODEPROMPT_INJECTION
Full Analysis
- [NO_CODE] (SAFE): The skill submission contains zero files. Without code or instructions, there is no active execution surface for threats.
- [PROMPT_INJECTION] (LOW): A vulnerability surface for indirect prompt injection was identified via the github.event.label.name variable mentioned in automated alerts. Evidence Chain: 1. Ingestion points: GitHub event label metadata. 2. Boundary markers: Not applicable/None. 3. Capability inventory: None. 4. Sanitization: None. This is a theoretical finding as no processing logic is present.
Recommendations
- Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata