apify-actorization

Fail

Audited by Snyk on Nov 19, 2026

Risk Level: CRITICAL
Full Analysis

CRITICAL E004: Prompt injection detected in skill instructions.

  • Potential prompt injection detected (medium risk: 0.30). The skill's content is an Apify Actorization guide with no evident intent to attack the user, but its Dockerfile example pipes a remote script directly to sh (curl ... | sh), which contradicts the skill's own security note against that practice and is a risky, unvetted install step.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (low risk: 0.10). The required workflow is mainly CLI setup and SDK wrapping, but the Crawlee examples and the test run (apify run --input '{"startUrl": ...}') have the agent drive a crawler against a user-supplied URL, so outsider web content is read only after the agent selects a URL.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

  • Potentially malicious external URL detected (medium risk: 0.30). The only questionable runtime fetch is the Dockerfile example in references/cli-actorization.md, which pipes the ubi bootstrap script from houseabsolute/ubi (an individual-owned repository, not first-party Apify) to sh during build. That is executed remote code from a common but unofficial source, so the tier is medium. The other URLs are docs or Apify-official sources, and the npm/pip installs are package names without locators.

Issues (3)

E004
CRITICAL

Prompt injection detected in skill instructions.

W011
MEDIUM

Third-party content exposure detected (indirect prompt injection risk).

W012
MEDIUM

Unverifiable external dependency detected (runtime URL that controls agent).

Audit Metadata
Risk Level
CRITICAL
Analyzed
Nov 19, 2026, 11:19 AM
Issues
3