jobs-evaluate
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download installation scripts from the developer's official repository (solid-company) on GitHub.
- [REMOTE_CODE_EXECUTION]: Provides commands to download and execute shell or PowerShell scripts directly from the developer's repository for the purpose of installing the sjctl utility.
- [COMMAND_EXECUTION]: Runs the sjctl CLI tool to retrieve JSON-formatted profiles and persist evaluation results; includes a fallback to run code via the Go compiler.
- [PROMPT_INJECTION]: Processes external data (job offers) and local profile data, which presents an attack surface for indirect prompt injection.
- Ingestion points: Cached job offer data and sjctl profile JSON output.
- Boundary markers: Uses a structured rubric for scoring and passes data to CLI tool arguments.
- Capability inventory: Executes local binaries (sjctl) and developer tools (Go).
- Sanitization: Data is handled through structured CLI flags and JSON parsing, limiting the impact of malicious content.
Audit Metadata