detection-sigma

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFENO_CODEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [NO_CODE]: The skill documentation and workflows reference several Python scripts (e.g., scripts/sigma_convert.py, scripts/sigma_validate.py, and scripts/compliance_coverage.py) that are not included in the provided file set. The skill consists entirely of markdown documentation and YAML rule templates.\n- [SAFE]: All identified dependencies and external references point to official, well-known security projects and repositories (e.g., SigmaHQ, pySigma, MITRE ATT&CK). No suspicious external communication or code execution was found.\n- [PROMPT_INJECTION]: The skill is designed to ingest third-party Sigma rules from a local directory (rules/). This creates an indirect prompt injection surface where malicious instructions could be embedded in the rule logic or metadata to influence the agent during processing.\n
  • Ingestion points: Reads Sigma rules from the rules/ directory for validation and conversion workflows.\n
  • Boundary markers: None identified in the provided templates or instructional text.\n
  • Capability inventory: Implies file read/write and network capabilities through its mentioned (but absent) scripts.\n
  • Sanitization: No input validation or sanitization of rule content is described.\n- [DATA_EXFILTRATION]: Example Python code provided in the documentation (references/backend-support.md) demonstrates disabling SSL certificate verification (verify=False) when interacting with SIEM APIs. While provided as a template, this is a discouraged security practice that could facilitate data interception if implemented without modification.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 11:00 AM
Security Audit — agent-trust-hub — detection-sigma