autosync-ic-skills

Pass

Audited by Gen Agent Trust Hub on Nov 19, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches a synchronization script and skill contents from https://skills.internetcomputer.org. These resources are hosted on the vendor's official infrastructure and are used as intended for the skill's auto-update feature.
  • [COMMAND_EXECUTION]: The skill and its associated script execute shell commands (curl, jq, rm, mv) to mirror remote skill files. The script implements safety functions is_safe_name and is_safe_relpath to validate remote inputs and prevent path traversal during file operations.
  • [PERSISTENCE]: Configures a SessionStart hook in .claude/settings.json to ensure the synchronization script runs automatically in future sessions. The documentation correctly notes that the platform will prompt the user to review and trust this hook before it becomes active.
  • [INDIRECT_PROMPT_INJECTION]: The skill synchronizes external SKILL.md files which are subsequently processed by the agent, creating a surface for indirect instructions.
  • Ingestion points: The script scripts/sync-ic-skills.sh downloads skill files into the .claude/skills/ directory.
  • Boundary markers: The skill does not apply explicit delimiters or warnings to the downloaded markdown files.
  • Capability inventory: The skill possesses file write, directory creation, and network download capabilities via curl and mv in scripts/sync-ic-skills.sh.
  • Sanitization: The script uses is_safe_name and is_safe_relpath to sanitize paths derived from the remote index, preventing unauthorized file system access.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 19, 2026, 11:00 AM