mirror of
https://github.com/magnus919/agent-skills.git
synced 2026-09-17 06:26:31 +03:00
Add a conditional skill for products with recurring human relationships. Covers success plans, health evidence, renewal/expansion signals, QBRs, handoffs, escalation, and closed-loop Voice of Customer. The skill is conditional: it declines and routes away when the product has no accounts, renewals, QBRs, or customer-success team. Includes: - SKILL.md with conditional frontmatter, trigger sections, four product- model adaptations (B2B subscription, transactional, public-service, internal product), core artifacts, privacy and human-judgment boundaries, and routing to product-analytics-and-measurement, product-adoption, product-experimentation, go-to-market, and product-lifecycle-learning. - README.md with all five required human-facing sections. - references/discovery-brief.md surveying existing content and defining ownership boundaries and routing. - references/privacy-and-human-judgment.md with consent framework, surveillance-risk guidance, decision-support rules, and data classification tiers. - templates/applicability-decision.md, templates/success-plan.md, templates/health-risk-record.md, and templates/escalation-and- feedback-closure.md. - evals/evals.json with 5 output-quality cases covering B2B subscription, internal-tool decline (negative trigger), public-service routing, renewal-risk with mixed signals, and conflicting health evidence. - Updated root README.md catalog section, references/skill-triggers.md, and regenerated marketplace/Codex/llms.txt catalogs. Co-authored-by: username <username> Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
6.7 KiB
6.7 KiB
Discovery Brief — Conditional Customer Success
Date: 2026-08-02
Scope: Issue #192 — feat: add conditional customer-success skill
Survey of Existing Content
What already exists in the catalog
| Existing Skill / Area | Relevant Content | Ownership Boundary |
|---|---|---|
go-to-market |
Account planning, customer segmentation, expansion strategy, pricing and packaging | Owns commercial motion (acquisition, conversion, pricing, account-tier design). CS consumes account-tier context; does not own commercial design. |
product-adoption |
Onboarding, activation, behavior change, feature discovery, sustained-use review, cohort segmentation, non-SaaS adoption contexts | Owns adoption-side diagnostics and plans. CS consumes adoption signals as health dimensions; does not own adoption methodology. |
product-analytics-and-measurement |
Metric trees, tracking plans, instrumentation QA, measurement governance | Owns metric definitions and instrumentation. CS consumes health metrics and evidence from the analytics layer. |
product-experimentation |
Experiment design, readout, ship/no-ship decisions, stopping rules | Owns experiment method. CS may propose experiments (intervention tests for at-risk accounts) but does not own experiment design. |
product-strategy |
North Star, product vision, strategic context, portfolio thinking | Owns strategic direction. CS informs strategy with customer evidence; does not set strategy. |
product-methodology |
Prioritization frameworks (RICE, etc.), discovery techniques, stakeholder alignment | Owns tactical product method. CS consumes prioritization context; does not own RICE or discovery. |
financial-modeling |
Unit economics, CAC, LTV, churn modeling, renewal forecasting | Owns financial model. CS provides account-level evidence (health, risk) that feeds financial models. |
data-scientist |
Statistical inference, causal analysis, experiment analysis | Owns statistical method. CS routes complex health-signal analysis here. |
product-roadmapping-and-portfolio |
Roadmap construction, portfolio balancing, investment allocation | Owns roadmap. CS feeds customer evidence into roadmap inputs; does not own prioritization. |
What is missing (the gap)
No dedicated customer-success capability exists in the catalog. The existing skills cover isolated pieces:
- Onboarding and activation (
product-adoption) but not ongoing success management or account health. - Renewal modeling (
financial-modeling) but not the human relationship practice (QBRs, success plans, escalation). - Expansion strategy (
go-to-market) but not the evidence-based account-expansion signal from CS. - Stakeholder alignment (
product-methodology) but not the closed-loop Voice of Customer process.
No skill currently defines:
- When customer-success practice applies and when it should not be loaded.
- Success plans as living artifacts connecting customer intent to product evidence.
- Health/risk records that are evidence-based, not automated scores.
- QBR structure and cadence.
- Escalation paths with explicit human-judgment gates.
- Handoff protocols between CS, product, support, and engineering.
- Closed-loop feedback closure (customer insight → product change → customer communication).
- Privacy boundaries specific to customer-success data.
Ownership Boundaries
What this skill owns
- Applicability decision: determining whether CS practice applies to a given product context.
- Success plans: desired outcomes, product-capability alignment, milestones, evidence gates, relationship ownership.
- Health/risk records: evidence-based, signal-per-dimension, with trend, confidence, and conflicting-signal tracking.
- QBR structure: evidence-review cadence and format (not sales presentations).
- Escalation paths: trigger thresholds, named decision-makers, human- judgment gates, fallback rules.
- Handoff protocols: structured handoffs to product, support, and engineering with account context and evidence.
- Closed-loop feedback: insight → validation → product decision → implementation → customer communication → closure record.
- Privacy and human-judgment boundaries: consent framework, surveillance-risk guidance, decision-support rules, escalation-gate requirements.
What this skill does NOT own (routes to specialists)
| Concern | Routed to | Why |
|---|---|---|
| Metric definition, tracking plans, instrumentation | product-analytics-and-measurement |
Owns measurement infrastructure |
| Adoption diagnostics, onboarding design, activation plans | product-adoption |
Owns adoption lifecycle |
| Statistical analysis of health signals | data-scientist |
Owns statistical method |
| Commercial pricing, packaging, contract terms | go-to-market |
Owns commercial design |
| Financial modeling (CAC, LTV, churn) | financial-modeling |
Owns financial models |
| Experiment design for CS interventions | product-experimentation |
Owns experiment method |
| Product roadmap decisions from CS evidence | product-roadmapping-and-portfolio |
Owns roadmap |
| Retirement communication, sunset coordination | product-lifecycle-learning |
Owns lifecycle closure |
Routing Decision Tree
Is there a recurring human relationship?
├── YES: Are there accounts, renewals, QBRs, OR a CS team?
│ ├── YES → LOAD conditional-customer-success
│ └── NO → Route to product-adoption (adoption-side) or product-lifecycle-learning
└── NO → DECLINE. Route to:
├── product-analytics-and-measurement (for health metrics)
├── product-adoption (for adoption diagnostics)
└── product-lifecycle-learning (for churn/retirement patterns)
Product-Model Applicability Matrix
| Model | Accounts? | Renewals? | QBRs? | CS Team? | CS Applies? | Adaptation |
|---|---|---|---|---|---|---|
| B2B Subscription | Yes | Yes | Yes | Yes | Full | Standard artifacts |
| Transactional | Partial | No (repeat purchase) | Event-triggered | Sometimes | Partial | Adapted success plans, purchase-frequency health |
| Public Service | Rare | No | Program-review | Rare | Minimal | Service-outcome plans; stricter privacy; often decline |
| Internal Product | Sometimes | No | Internal-service-review | Sometimes | Conditional | Internal SLA plans; often decline |
Routing to Skills Not Yet Landed
product-lifecycle-learning (prose reference) is the destination for:
- Sustained health decline after 2+ remediation attempts.
- Churn-pattern learning across the portfolio.
- Retirement communication plans and migration-support coordination.
Until it lands, this skill records the routing in prose. The routing becomes a resolved link when the directory exists.