code-annotation-patterns
Pass
Audited by Gen Agent Trust Hub on Nov 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill introduces structured annotation patterns (such as @ai-security, @ai-tech-debt, and @ai-change-impact) that an agent is expected to parse and potentially act upon. This design allows for instructions to be embedded in data (source code) that could influence the agent's actions if it automates tasks based on these markers.
- Ingestion points: Source code files processed by the agent (e.g., TypeScript, Python, Go, Java, and Rust files as shown in the examples).
- Boundary markers: Absent; the skill provides no mechanisms or instructions for the agent to verify the authenticity of annotations or to distinguish between benign developer comments and malicious injections.
- Capability inventory: The execution environment includes powerful tools such as Bash, Write, and Edit, which could be abused if an agent follows a malicious 'mitigation' or 'action' instruction found in an annotation.
- Sanitization: Absent; there is no requirement for content validation, escaping, or integrity checking for the metadata payload.
Audit Metadata