skills/mukul975/anthropic-cybersecurity-skills/implementing-vulnerability-sla-breach-alerting/Gen Agent Trust Hub
implementing-vulnerability-sla-breach-alerting
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The system processes external vulnerability data (JSON and CSV findings) and interpolates fields like CVE titles, asset hostnames, and owner information directly into notification templates and SQL queries. This creates a surface for indirect prompt injection or message spoofing if an attacker can control the content of vulnerability findings in the source platform.
- Ingestion points:
scripts/agent.py(loads JSON vulnerability data) andscripts/process.py(imports findings from CSV files). - Boundary markers: Not implemented in the provided alerting logic or notification templates.
- Capability inventory: The skill possesses the capability to write data to Slack/Teams webhooks, trigger PagerDuty incidents, create Jira tickets via API, and send emails via SMTP.
- Sanitization: Absent. The scripts directly use fields from the ingested data in alert payloads and email bodies, which could lead to message spoofing or HTML injection in email clients if the source data contains malicious payloads.
Audit Metadata