Skills
Navigate
skills/lassejlv/skills/gcp-account-cleanup/Gen Agent Trust Hub

gcp-account-cleanup

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill leverages the gcloud CLI to perform resource discovery and deletion. These actions are strictly controlled by a 'Non-Negotiable Safety Flow' which requires two distinct human confirmation steps and a mandatory dry-run summary before any destructive commands are issued.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from GCP API outputs (resource names, tags, and metadata) which are then processed by the agent.
  • Ingestion points: references/gcp-inventory.md via various gcloud ... list commands.
  • Boundary markers: Instructions mandate specific verbatim confirmation phrases and dry-run summaries to be presented to the user.
  • Capability inventory: gcloud resource deletion and management capabilities described in SKILL.md and references/gcp-inventory.md.
  • Sanitization: The safety flow requires human validation of the dry-run summary, providing a manual check against potentially malicious metadata injected via resource names.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 09:55 PM