django-celery
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill attempts to ingest project-specific information from
.agents/django-project-context.mdto configure the broker and deployment setup. Malicious instructions within this file could potentially manipulate the agent's logic or behavior when providing setup advice. - Ingestion points:
.agents/django-project-context.md(read by the agent at initialization) - Boundary markers: Absent
- Capability inventory: Execution of background worker commands, configuration of environment variables, and generation of integration code
- Sanitization: Absent
- [INDIRECT_PROMPT_INJECTION]: The skill provides code patterns for task chains that fetch data from external sources (e.g.,
fetch_data.s(source_url)). Data retrieved from these untrusted external locations could contain prompt injections designed to influence subsequent steps in the task pipeline. - Ingestion points: Remote data sources targeted by Celery tasks
- Boundary markers: Absent
- Capability inventory: Network requests, database writes, and external API calls via Celery tasks
- Sanitization: Absent
Audit Metadata