coding-cli

Pass

Audited by Gen Agent Trust Hub on Apr 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands to manage a workspace environment within the /root/hiclaw-fs/ directory, including directory creation (mkdir), file copying (cp), and navigation (cd).
  • [EXTERNAL_DOWNLOADS]: The skill employs git clone to fetch source code from external repositories as part of the workspace preparation process.
  • [DATA_EXFILTRATION]: The skill uses the mc mirror command to synchronize data from the local filesystem to an external storage service defined by the ${HICLAW_STORAGE_PREFIX} environment variable. This is used to share the workspace with the Manager component.
  • [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it processes task requirements from an external spec.md file and incorporates them into a prompt generated for downstream AI tools like Claude Code or Gemini CLI.
  • Ingestion points: The skill reads task instructions and coding requirements from a spec.md file.
  • Boundary markers: None identified; the generated prompt directly includes task requirements.
  • Capability inventory: The skill can execute shell commands, clone repositories, and synchronize files to remote storage.
  • Sanitization: There is no explicit validation or sanitization of the input task description before it is interpolated into the prompt for the CLI tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 6, 2026, 08:11 AM