ln-404-test-executor

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill reads 'docs/project/infrastructure.md' to access sensitive configuration data including server IPs, ports, and service endpoints.
  • [COMMAND_EXECUTION]: The skill executes test-related shell commands and 'Docker commands' as prescribed in the project's 'docs/project/runbook.md'.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted data from task providers which then influences the execution of shell commands and file operations. • Ingestion points: Task descriptions from Linear and local files. • Boundary markers: None identified. • Capability inventory: File system read/write/edit, Linear API access, and shell execution. • Sanitization: No sanitization of task content is performed prior to processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 07:47 AM