django-models

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by instructing the agent to ingest external data from project context files.\n
  • Ingestion points: Reads .agents/django-project-context.md and .claude/django-project-context.md as specified in SKILL.md.\n
  • Boundary markers: Absent; there are no instructions to delimit or ignore instructions within these context files.\n
  • Capability inventory: No executable scripts are included in the skill; the skill body provides instructions for the agent to generate and present shell commands and Python code for model definition.\n
  • Sanitization: Absent; no validation or filtering is specified for the ingested context data.\n- [COMMAND_EXECUTION]: The skill documents standard Django management commands such as makemigrations, migrate, and sqlmigrate. These are intended for manual project administration and are not executed automatically by the skill.\n- [NO_CODE]: The skill consists entirely of Markdown documentation and reference materials, without any executable scripts, binary files, or automated installation processes.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 09:06 PM