skills/mukul975/anthropic-cybersecurity-skills/implementing-vulnerability-remediation-sla/Gen Agent Trust Hub
implementing-vulnerability-remediation-sla
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The reporting engine in
scripts/process.pyandscripts/agent.pyprocesses external vulnerability data to generate HTML dashboards without sanitizing the input. - Ingestion points:
scripts/agent.py(line 35) andscripts/process.py(lines 218-219) read data from JSON and CSV files. - Boundary markers: The scripts lack delimiters or instructions to ignore malicious content within the ingested data.
- Capability inventory: The skill has the capability to write files to the local system, specifically generating HTML report files.
- Sanitization: The HTML generation in
scripts/process.py(line 207) uses direct string interpolation of fields likehostnameandcvewithout escaping, which could allow for cross-site scripting (XSS) if the input data is compromised. - [EXTERNAL_DOWNLOADS]: The skill includes documentation for integrating with Tenable's vulnerability management platform.
- The
references/api-reference.mdfile provides code snippets that interact withcloud.tenable.comto retrieve vulnerability workbenches. - These operations are standard for security workflows and utilize environment variables for secure credential handling.
Audit Metadata