session-scout

Pass

Audited by Gen Agent Trust Hub on Apr 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses PowerShell on Windows (Session-Scout.ps1) and Bash on Unix (session-scout.sh) to execute system commands including find, stat, grep, and wsl.exe. These are used to locate and analyze session files in various directories.- [COMMAND_EXECUTION]: Instructions in SKILL.md recommend running PowerShell with -ExecutionPolicy Bypass, which lowers environment security to allow the execution of unsigned scripts.- [DATA_EXFILTRATION]: The skill scans sensitive directories containing personal application data, such as ~/.claude/projects, %APPDATA%\Claude, and ~/.local/share/opencode/log. These files contain session transcripts, metadata, and history. The skill reads portions of these files to extract project paths and provides functionality to export the results to a local CSV file.- [PROMPT_INJECTION]: The skill presents an Indirect Prompt Injection surface by processing untrusted data from session logs:
  • Ingestion points: Content is read from .jsonl and .log files in application data directories (Session-Scout.ps1, session-scout.sh).
  • Boundary markers: Absent. The scripts do not use delimiters or instructions to ignore embedded content when extracting data.
  • Capability inventory: Local file reading, CSV file writing, and execution of shell commands/WSL.
  • Sanitization: Absent. The scripts use regex to extract specific metadata fields but do not validate or sanitize the input strings before presenting them to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 10, 2026, 08:14 AM
Security Audit — agent-trust-hub — session-scout