Files
gh-gtmagents-gtm-agents-plu…/commands/configure-rules.md
2025-11-29 18:31:13 +08:00

50 lines
3.3 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
---
name: configure-rules
description: Deploys decision logic, content variants, and delivery rules across personalization channels.
usage: /personalization-engine:configure-rules --initiative "PLG Onboarding" --environment staging --channels "web,in-app"
---
# Command: configure-rules
## Inputs
- **initiative** reference to the personalization effort from `define-profiles`.
- **environment** staging | production to govern deployment steps.
- **channels** comma-separated list of activation surfaces.
- **change_type** net-new, update, rollback.
- **approvers** optional stakeholders for governance sign-off.
### GTM Agents Pattern & Plan Checklist
> Mirrors GTM Agents orchestrator blueprint @puerto/plugins/orchestrator/README.md#112-325.
- **Pattern selection**: Rule configuration generally runs **pipeline** (pre-flight → decision build → variant mapping → QA → deployment). If decision build + variant prep happen in parallel, note a **diamond** block with merge gate in the plan header.
- **Plan schema**: Save `.claude/plans/plan-<timestamp>.json` capturing initiative, environments, dependency graph (data eng, creative, QA, governance), error handling, and success metrics (latency, personalization lift, incident count).
- **Tool hooks**: Reference `docs/gtm-essentials.md` stack—Serena for rule diffing, Context7 for platform SOPs, Sequential Thinking for go/no-go reviews, Playwright for simulation/QA evidence capture.
- **Guardrails**: Default retry limit = 2 for deployment/QA failures; escalation ladder = Personalization Architect → Data Privacy Lead → Exec sponsor.
- **Review**: Run `docs/usage-guide.md#orchestration-best-practices-puerto-parity` before deployment to confirm dependencies + approvals.
## Workflow
1. **Pre-flight Review** validate profiles, data freshness, consent status, and experiment dependencies.
2. **Decision Flow Build** configure rules, weights, or model endpoints in MAP/CDP/product tooling.
3. **Variant Mapping** link each rule outcome to content assets, CTAs, and fallback experiences.
4. **QA & Simulation** run synthetic traffic through decision trees, capture screenshots/logs.
5. **Deployment & Logging** push changes via API/CLI, note version metadata, set up monitoring hooks.
## Outputs
- Deployment runbook with rule IDs, version numbers, and rollback plan.
- QA evidence (simulation results, screenshots, payload logs).
- Governance log including approvers, timestamps, and linked experiments.
- Plan JSON entry stored/updated in `.claude/plans` for audit trail.
## Agent/Skill Invocations
- `customer-data-engineer` ensures data pipelines and environments are ready.
- `personalization-architect` verifies experience logic + content mapping.
- `content-variants` skill tracks asset requirements + approvals.
- `governance` skill enforces change controls and compliance steps.
## GTM Agents Safeguards
- **Fallback agents**: document substitutes (e.g., governance lead covering architect) when owners unavailable.
- **Escalation triggers**: if QA fails twice, latency spikes, or privacy gate blocks deployment, trigger GTM Agents rip-cord and log remediation in plan JSON.
- **Plan maintenance**: update plan JSON/change log when rule sets, environments, or deployment windows change so reviewers can trace history.
---