skills/nasrulhazim/claude/package-dev/Gen Agent Trust Hub

package-dev

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches boilerplate package templates from Spatie's GitHub repositories during the project scaffolding process.\n- [COMMAND_EXECUTION]: Executes shell commands for git operations (clone, tag, push), dependency management (composer update), and network requests (curl to Packagist API).\n- [COMMAND_EXECUTION]: Triggers the execution of the package's test suite via composer test or direct calls to the Pest binary.\n- [REMOTE_CODE_EXECUTION]: Automatically runs a configure script included in the third-party skeleton templates after cloning them.\n- [PROMPT_INJECTION]: Identifies an indirect prompt injection surface where user-supplied strings are interpolated into the generated package files and configuration.\n
  • Ingestion points: User-provided vendor name, package name, and description in the scaffold command within SKILL.md.\n
  • Boundary markers: No specific delimiters or safety instructions are defined for the interpolation of user input.\n
  • Capability inventory: Includes file system writes, network requests via curl, and shell command execution (git, composer) as described in SKILL.md.\n
  • Sanitization: The skill does not implement explicit sanitization or validation for user-provided metadata.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 11:55 PM
Security Audit — agent-trust-hub — package-dev