rch

Pass

Audited by Gen Agent Trust Hub on Nov 2, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPERSISTENCEDATA_EXFILTRATIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill's primary purpose is the remote execution of compilation commands on worker machines to offload local CPU/memory pressure. This behavior is intentional and documented as the core functionality.
  • [COMMAND_EXECUTION]: Provides extensive capabilities for the agent to run diagnostic and administrative commands on remote workers, including disk usage checks (df), process monitoring (lsof), and permission management (chown/chmod).
  • [PERSISTENCE]: Installs long-lived interception mechanisms, including a PreToolUse hook in agent settings (e.g., ~/.claude/settings.json) and a toolchain shim in ~/.rch/shims/cargo, to ensure build commands are consistently routed to the offload daemon.
  • [DATA_EXFILTRATION]: Syncs the local workspace to remote workers for building. To protect user data, the skill defaults to excluding sensitive files such as .env, .key, and secrets.json from these transfers.
  • [CREDENTIALS_UNSAFE]: Interacts with local SSH identity files (e.g., ~/.ssh/id_ed25519) to manage authentication with the remote worker fleet.
  • [EXTERNAL_DOWNLOADS]: Interacts with the author's official GitHub repository (github.com/Dicklesworthstone/remote_compilation_helper) for source code and binary management. These represent legitimate vendor resources within the context of the skill author.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 2, 2026, 03:43 AM