skills/mukul975/anthropic-cybersecurity-skills/performing-dmarc-policy-enforcement-rollout/Gen Agent Trust Hub
performing-dmarc-policy-enforcement-rollout
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes DMARC aggregate reports in XML format from external sources.
- Ingestion points: The
parse_dmarc_reportfunction inscripts/process.pyreads and parses XML files provided as input. - Boundary markers: No delimiters or instructions are provided to the agent to ignore potentially malicious embedded content within the reports.
- Capability inventory: The skill uses
dnspythonfor network DNS lookups and has file-reading capabilities. - Sanitization: The implementation uses standard
xml.etree.ElementTreefor parsing without specific mitigations for XML-based vulnerabilities. - [METADATA_POISONING]: The skill contains inconsistent author information. The YAML frontmatter in
SKILL.mdlists the author as 'mahipal', while theLICENSEfile attributes the work to 'mukul975'.
Audit Metadata