stack-trace-decoder
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted stack traces provided by the user, which presents a surface for indirect instructions.
- Ingestion points: Users are prompted to paste stack traces directly into the agent's context within the 'AI-Assisted Decoding' section of
SKILL.md. - Boundary markers: The skill suggests using triple backticks as a structural delimiter for input, which provides minimal isolation.
- Capability inventory: The skill contains analysis and parsing logic; it does not include instructions for dangerous tool execution, system persistence, or network exfiltration.
- Sanitization: There is no evidence of sanitization, validation, or filtering applied to the input stack trace data.
Audit Metadata