axiom
Pass
Audited by Gen Agent Trust Hub on Apr 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references and provides installation commands for several official Axiom SDKs and tools from well-known registries and official sources.
- Fetches the Axiom CLI via Homebrew using the official tap (
axiomhq/tap/axiom). - Installs Node.js integration packages (
@axiomhq/js,@axiomhq/pino,next-axiom,@axiomhq/web-vitals) from the public npm registry. - Installs the Python SDK (
axiom-py) from PyPI. - [PROMPT_INJECTION]: The skill describes a workflow where the agent analyzes user-provided logs to troubleshoot issues, creating a surface for indirect prompt injection.
- Ingestion points: User-provided logs and error outputs in troubleshooting scenarios (Example 2 in
SKILL.md). - Boundary markers: No explicit markers or delimiters are defined to separate untrusted log data from agent instructions.
- Capability inventory: The skill suggests the agent can modify configuration files and adjust system resource limits to apply fixes.
- Sanitization: No sanitization or validation of the input logs is mentioned.
Audit Metadata