kata-neckbeard
Warn
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: MEDIUMMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [METADATA_POISONING]: The skill metadata includes an 'optimized_for' field referencing 'gpt-6-astra', which is a non-existent model version. This provides misleading information regarding the skill's testing environment and its intended capabilities.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted content from repository files, which creates a potential surface for indirect prompt injection.
- Ingestion points: Source code, tests, configuration files, and documentation within the target repository (SKILL.md).
- Boundary markers: The instructions lack specific boundary markers or explicit directives for the agent to ignore instructions embedded within the data it analyzes.
- Capability inventory: The skill restricts the agent to investigation and reporting, with explicit instructions not to edit files or change code, and to preserve Git state (SKILL.md).
- Sanitization: No sanitization or filtering mechanisms are defined for the repository content processed by the agent.
Audit Metadata