arch-linux-triage
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input which can contain malicious instructions intended to manipulate the agent's output into proposing destructive commands.
- Ingestion points: Untrusted inputs enter the context via
${input:ArchSnapshot},${input:ProblemSummary}, and${input:Constraints}inSKILL.md. - Boundary markers: Absent. The instructions do not use specific delimiters or enclose inputs to prevent the model from executing text-based instructions embedded within the inputs.
- Capability inventory: The skill does not automate command execution or invoke tools directly; it generates text and copy-paste commands intended for manual execution by the user.
- Sanitization: Absent. No filtering, validation, or escaping rules are defined for the input parameters.
Audit Metadata