supabase
Fail
Audited by Snyk on Nov 13, 2026
Risk Level: CRITICAL
Full Analysis
CRITICAL E004: Prompt injection detected in skill instructions.
- Potential prompt injection detected (medium risk: 0.30). The skill is a legitimate Supabase toolkit with no sign of attacking the user, but its documented exec_sql helper is a SECURITY DEFINER function that runs arbitrary SQL and would be exposed through PostgREST, weakening database protections beyond what the advertised task needs.
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (low risk: 0.10). The required workflow runs agent-chosen SQL and REST reads against the user's own Supabase tables, so any outsider-submitted text stored in those rows (e.g., signup or comment fields) reaches the LLM only after the agent selects a specific table or query, not through a queue the outsider can force into the workflow.
MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).
- Potentially malicious external URL detected (medium risk: 0.30). The README's documented install step runs
git cloneof the skill's own GitHub repository into the skills directory, so installed skill material is inbound from an individual-owned repository (not a recognized vendor), with no concrete warning signs, making it common/reputable unofficial and therefore medium; the OpenAI and Supabase API calls are outbound data exchanges and the npm package name lacks a URL, so they do not count.
Issues (3)
E004
CRITICALPrompt injection detected in skill instructions.
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
W012
MEDIUMUnverifiable external dependency detected (runtime URL that controls agent).
Audit Metadata