home/coding/agent-workflow-instructions

Agent Workflow.instructions

GPTClaudeDeepSeek··▲ 1,295 copies·updated 2026-07-14

agent-workflow-instructions.prompt

---
applyTo: "**"
---

Follow AGENTS.md as the source of truth.

Required load order:

1. AGENTS.md
2. agent_docs/active_context.md
3. only specific agent_docs/\*.md needed for the task

Default execution chain:
prompt-engineer -> owner(s) -> tester -> reviewer

Escalation:

- Include security for auth, permissions, secrets, uploads, risky integrations, or sensitive data.
- Include legal for privacy, consent, analytics, retention, AI disclosures, or policy-sensitive copy.
- Include devops for CI/CD, deployment, environment, and infrastructure delivery changes.

Handoff contract (required between agents):

- goal
- scope (files/symbols/flows)
- changes
- evidence (commands/tests/results)
- risks (assumptions/open questions)
- next-owner
- done-criteria

Definition of done:

- owning layer change is correct
- runtime behavior validated when relevant
- cross-layer contract mismatches checked
- reviewer-level risks acceptable

when to use it

Community prompt sourced from the open-source GitHub repo YoussefSelk/VibeCode-PlayBook (MIT). A "Agent Workflow.instructions" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.