automation-scripts

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The deployment script (scripts/deploy.sh) in SKILL.md ingests untrusted data via command-line arguments for the environment and version. While the environment is whitelisted, the version string is interpolated directly into docker, kubectl, and curl commands without sanitization, creating a potential surface for injection if the version source (e.g., git tags or CI inputs) is compromised.
  • Ingestion points: Command-line arguments $1 (ENVIRONMENT) and $2 (VERSION) in the scripts/deploy.sh section of SKILL.md.
  • Boundary markers: Absent for the interpolated shell variables.
  • Capability inventory: The script utilizes docker (build, tag, push), kubectl (set image, rollout status), and curl (smoke tests, Slack notifications).
  • Sanitization: The ENVIRONMENT variable is validated against a fixed list (staging|production), but the VERSION variable lacks escaping or validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 09:40 PM