Skip to content

Latest commit

 

History

History
54 lines (36 loc) · 2.06 KB

File metadata and controls

54 lines (36 loc) · 2.06 KB

Agent entry point

Verify first. Fix only if confirmed.

This file is the short cross-agent entry point for TradingChassis infrastructure work. It is not a second copy of the Cursor rules.

Canonical rules

Durable project invariants: .cursor/rules/safe-infrastructure-development.mdc

Implementation tasks: .cursor/rules/implementation-workflow.mdc

Independent reviews: .cursor/rules/review-workflow.mdc

Human-facing workflow: docs/AI_AGENT_WORKFLOW.md

Language

  • Respond in the language explicitly requested by the current task or prompt.
  • If no response language is specified, use the language of the user's current request.
  • Keep repository content in English unless the task explicitly requires otherwise.

Hard limits

  • Do not perform live OCI, Terraform, Ansible, Kubernetes, Helm, MicroK8s, host-network, firewall, storage, package, or service operations.
  • Do not read or output real secrets, credentials, private keys, Terraform state, tfvars, kubeconfigs, or vault material.
  • Repository guardrails are not a security boundary and do not prove sandbox isolation.

Ownership (summary)

  • Terraform owns OCI resources.
  • Ansible owns host configuration, storage, MicroK8s, and initial Argo CD bootstrap.
  • Argo CD owns long-lived Kubernetes resources.

Git authorization

  • Never work directly on main.
  • One logical change per branch and pull request.
  • Commit, push, pull-request creation, and merge only when the current task explicitly authorizes that action.
  • An implementation task is not permission to commit, push, or merge.

Safe wrappers

  • ./tools/check-agent-safety — local, read-only safety self-check
  • ./tools/validate-safe — local static validation wrapper
  • ./tools/check-sensitive-metadata — operator/live metadata hygiene (not a credential scanner)

Credential scanning runs in CI via pinned Gitleaks. See docs/REPOSITORY_SECURITY.md.