/security-flow-authorize

name: security-flow-authorize
description: "Phase 2 Authorize of security-flow"
disable-model-invocation: true
user-invocable: false
alwaysApply: false
tags: ["security", "workflow-phase"]
baseSchema: docs/schemas/phase.md

<security_flow_authorize>

<description_and_purpose> Recommend an enterprise-safe run contract for user approval or amendment. </description_and_purpose>

<workflow_context> Phase 2. Requires a passing or user-approved DEV/QA-envs secret gate. </workflow_context>

<phase_steps> 1. Assess contextual risk 2. Recommend run contract 3. Explain material tradeoffs 4. Surface decisions for approval </phase_steps>

<recommend_contract step="2.1" subagent="engineer" role="Enterprise security scope and authorization advisor" subagent_required_model="Claude Sonnet 5, GPT-5.6 Terra, Gemini 3.7 Flash">

1. USE SKILL `subagent-directives`. 2. USE SKILL `security`. 3. USE SKILL `risk-assessment`. 4. Recommend targets, environment, exclusions, audience, retention, and exploit-detail policy. 5. Recommend activities, tools, credentials, data flows, limits, and stop conditions. 6. Recommend active-test identities, routes, methods, rates, payloads, duration, and cleanup where applicable. 7. Explain coverage gained, residual gaps, and enterprise tradeoffs. 8. Mark each activity local read-only, separately gated, or prohibited. 9. Update `security-flow-state.md`.

</recommend_contract>

<handoff step="2.2"> Return the recommended contract with every material decision listed as awaiting approval. Raise material unknowns only; do not nitpick. The orchestrator runs the approval gate with the user. </handoff>

<validation_checklist>

• Approval names target and environment.
• Exclusions and stop conditions are explicit.
• External data flows are separately decided.
• Production active testing is absent.

</validation_checklist>

</security_flow_authorize>