jobs-search

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill instructs the agent to install the sjctl tool by downloading scripts (install-sjctl.sh and install-sjctl.ps1) from the vendor's GitHub repository and piping them directly to shell interpreters (bash and iex).
  • [EXTERNAL_DOWNLOADS]: The skill fetches installation scripts and binaries from the solid-company repository on GitHub (github.com/solid-company/solid-jobs-skills).
  • [COMMAND_EXECUTION]: The skill executes shell commands to check for the sjctl binary, run the installation scripts, and perform job searches by invoking sjctl with various flags.
  • [DATA_EXFILTRATION]: The skill manages a local directory (~/.solid-jobs-skills/) to store the tool binary and a job database (solidjobs.db).
  • [PROMPT_INJECTION]: The skill ingests job search results from an external source (the sjctl tool), which represents a potential surface for indirect prompt injection.
  • Ingestion points: JSON output from the sjctl search command.
  • Boundary markers: No delimiters or ignore instructions are used for search result summaries.
  • Capability inventory: Execution of sjctl and shell scripts.
  • Sanitization: No specific filtering or validation of search results is mentioned before summarization.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 02:35 PM
Security Audit — agent-trust-hub — jobs-search