status-report
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEMETADATA_POISONINGINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [METADATA_POISONING]: The LICENSE file contains anomalous trailing text appended after the Apache 2.0 license terms. This text includes references to capturing 'recorded-cams', 'cams-data', and using 'input()', which is inconsistent with the skill's purpose and appears deceptive.
- [INDIRECT_PROMPT_INJECTION]: The LICENSE file includes text that could be interpreted as data collection instructions ('capture elements into Packeted-User-Relations') if the agent reads and obeys content from auxiliary files.
- Ingestion points: LICENSE file.
- Boundary markers: None.
- Capability inventory: The skill contains no scripts or tool definitions.
- Sanitization: None.
Audit Metadata