buddy-reroll

Fail

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: HIGHCOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill reads the user's Claude Code configuration file located at ~/.claude.json to retrieve the accountUuid or userID. This sensitive information is required to replicate the pet generation logic locally.
  • [COMMAND_EXECUTION]: The patcher.py script performs binary patching on the Claude Code installation (e.g., cli.mjs or claude.exe) by replacing original salt bytes with custom-calculated ones. This involves direct modification of application code on the user's system.
  • [EXTERNAL_DOWNLOADS]: SKILL.md includes instructions to download the Bun runtime installer from its official domain (bun.sh) to facilitate high-performance hashing.
  • [REMOTE_CODE_EXECUTION]: The recommended installation method for the Bun runtime involves piping a remote script from https://bun.sh/install directly into a shell (bash), which is a high-risk pattern for executing untrusted remote code.
  • [COMMAND_EXECUTION]: The skill generates a Javascript snippet at runtime and executes it via the Bun process (bun -e) to perform the high-volume hashing needed for the brute-force search.
  • [COMMAND_EXECUTION]: Scripts use subprocess.run and subprocess.Popen to call system utilities such as where and which to locate application binaries across different operating systems.
Recommendations
  • HIGH: Downloads and executes remote code from: https://bun.sh/install - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Apr 7, 2026, 05:54 AM
Security Audit — agent-trust-hub — buddy-reroll