adr-decision-extraction
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to extract architectural decisions from external, untrusted sources such as conversation transcripts and design discussions without utilizing explicit boundary markers or input sanitization.\n
- Ingestion points: Processes user-provided conversation history and transcripts as defined in the extraction rules (SKILL.md).\n
- Boundary markers: No delimiters or specific instructions to ignore embedded commands are implemented to protect the agent from instructions hidden within the processed text.\n
- Capability inventory: The skill is restricted to structured JSON data extraction and lacks capabilities for network access, file system modification, or command execution.\n
- Sanitization: Input content is processed without filtering, validation, or escaping.
Audit Metadata