axiom-analyze-test-failures
Pass
Audited by Gen Agent Trust Hub on Apr 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to facilitate test analysis and crash diagnosis. Evidence in
SKILL.mdincludesls -lt ~/Library/Logs/DiagnosticReports/*.ips,xcsym crash, andswift test. These commands are used to locate crash reports, symbolicate them, and reproduce failures through test iterations. - [DATA_EXFILTRATION]: The skill accesses diagnostic information from system crash reports. Evidence in
SKILL.mdshows instructions to read.ipsfiles from~/Library/Logs/DiagnosticReports/. This access is necessary for 'Pattern 6: Test-generated crashes' to determine the root cause of process failures. No evidence of network exfiltration was detected. - [PROMPT_INJECTION]: The skill analyzes external content such as Swift source code and crash logs, creating a surface for indirect prompt injection. 1. Ingestion points: Swift source files and crash report files specified in
SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: Shell command execution viaswift testandxcsymdocumented inSKILL.md. 4. Sanitization: Absent. The risk is considered low as the operations are confined to a local developer environment and are standard for diagnostic workflows.
Audit Metadata