All skills
simota avatar

/ripple

@c805268
by shingo imotasimota/agent-skills85 stars
15

Analyzing pre-change impact across vertical (dependency chains, files) and horizontal (pattern consistency, naming) dimensions. Use to estimate blast radius before a refactor. No code.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/ripple

This session only. Nothing lands on disk.

referenceblast-radius-quant.md

≈2.2k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Blast Radius Quantification Reference

Purpose: Quantify the production blast radius of a proposed change — customer count, SLO error-budget burn, revenue at risk, region/AZ scope, tenant fan-out, and data classification. Produces an evidence-grounded SEV-tier mapping that translates code-level change impact into incident-severity language operators and leadership understand.

Scope Boundary

  • ripple blast-radius: Production-quantified impact — customers, revenue, SLO burn, scope, data class, SEV tier (this document).
  • ripple impact / vertical / horizontal (elsewhere): Code-level dependency tracing. Feed dependency counts into blast-radius quantification.
  • ripple rollback-plan (elsewhere): Reversibility contract. Blast radius informs rollback urgency (ToR target).
  • Beacon (elsewhere): SLO/SLI design and error-budget policy. Blast-radius quantification consumes the SLO/error-budget from Beacon.
  • Triage (elsewhere): Live incident scoping. Blast-radius here is pre-change estimation; Triage is in-incident actual.
  • Sentinel (elsewhere): Security blast radius (exploit reach). Coordinate when change touches auth/PII/payment paths.

Workflow

SCOPE      →  confirm change target + dependency count (from `impact`)
           →  identify surfaces: API, worker, DB, cache, queue, UI

CUSTOMERS  →  count active customers / users through affected surface
           →  segment by plan tier (enterprise / paid / free)
           →  flag if any one tenant > 5% of traffic (Noisy Neighbor risk)

REVENUE    →  compute revenue-at-risk (MRR through surface × downtime window)
           →  pair with SLA contract obligations (% uptime commitments)

SLO BURN   →  map to error-budget burn rate
           →  e.g., 1% of month's budget = SEV3 threshold

REGION/AZ  →  single-AZ / multi-AZ / multi-region
           →  mark blast floor (minimum recoverable scope)

TENANT     →  multi-tenant fan-out factor
           →  RLS/isolation verification (defers to Schema `tenant`)

DATA CLASS →  PII / PHI / financial / IP / internal
           →  activate compliance blast overlay (GDPR/HIPAA/PCI)

SEV MAP    →  combine dimensions → SEV1/SEV2/SEV3/SEV4
           →  emit escalation triggers

Blast Dimensions

Dimension Measure Data Source
Customer count # active accounts through the affected surface in last 28 days analytics warehouse / billing DB
User count # DAU / MAU reaching the affected code path product analytics
Revenue at risk MRR × (expected downtime minutes / month minutes) billing DB × on-call history
SLA commitment contractual uptime % (99.9 / 99.95 / 99.99) sales / legal
Error budget remaining minutes in current window Beacon SLO dashboard
Region scope single-AZ / multi-AZ / multi-region infra registry
Tenant fan-out % of tenants touched + largest-tenant share tenant routing table
Data classification PII / PHI / Financial / IP / Internal data catalog
Dependent services # downstream services consuming the surface service catalog / Atlas output
AI-assisted yes/no (elevated scrutiny per Amazon 2026) PR metadata

Customer and Revenue Formula

Customers Affected       = distinct_accounts(events ∋ affected_endpoint, last 28d)
Revenue at Risk (monthly)= Σ MRR(customer) for customers_affected
Expected Impact ($)      = Revenue_at_Risk × (expected_downtime_min / 43200)
                           (43200 = minutes/month)
SLA Breach Risk          = YES if expected_downtime_min > monthly_budget
                           e.g., 99.9% budget = 43.2 min/month

SEV Tier Mapping

SEV Condition On-call Response Examples
SEV1 Customer-visible outage + >50% MAU or >25% revenue affected OR PII/PHI/Financial data at risk Immediate page, exec notification, customer comms within 30 min Payment processor down, auth system unavailable, PII leak
SEV2 Major degradation + 10-50% MAU or top-tier enterprise tenant affected OR SLO burn > 2x budget Page within 15 min, status page update One region down, feature broken for paid tier
SEV3 Partial degradation + < 10% MAU, no enterprise tenant, SLO burn within budget Page during business hours, team-level comms Secondary feature degraded, one AZ impaired
SEV4 Minor issue, internal impact only, no customer-visible symptom Ticket, no page Internal dashboard broken, dev-only surface

Compliance Blast Overlay

When the change touches a classified data path, add compliance blast estimates:

Classification Regulation Added Blast
PII (EU residents) GDPR Data breach notification obligation (72h), fine up to 4% global revenue or €20M
PII (CA residents) CCPA/CPRA Consumer notification, statutory damages $100-$750 per consumer
PII (JP residents) APPI PPC notification, 三層分類 handling
PHI HIPAA HHS breach notification, per-record penalties
Financial (PAN) PCI-DSS Card brand notification, forensic audit
Financial (SOX scope) SOX 10-K/10-Q disclosure impact

Route compliance-impacting changes to Cloak and Canon[regulatory] for parallel review.

Output Template

## Blast Radius Report

### Change Summary
- **Target**: [file paths / API surface]
- **Type**: [schema / API / worker / UI / infra]
- **Deploy surface**: [production / staging / preview]

### Quantified Blast

| Dimension | Value | Source |
|-----------|-------|--------|
| Customers affected | [N] | [source] |
| % of MAU | [X%] | product analytics |
| Revenue at risk (mo) | $[X] | billing |
| Top-tier tenants affected | [list] | CRM |
| Region scope | [AZ/region list] | infra registry |
| Tenant fan-out | [X% of tenants, largest X%] | tenant routing |
| Data classification | [PII/PHI/Financial/IP/Internal] | data catalog |
| Dependent services | [list] | service catalog |

### SLO Impact
- **Affected SLO**: [name, target %]
- **Remaining error budget**: [minutes / %]
- **Expected burn from this change**: [minutes / %]
- **Burn classification**: within / 2x / exhausted

### SEV Tier (Pre-Change Estimate)
- **Tier**: [SEV1 / SEV2 / SEV3 / SEV4]
- **Rationale**: [which dimensions drove the classification]
- **Escalation triggers**: [conditions under which tier escalates during rollout]

### Compliance Overlay
- **Regulations in scope**: [GDPR / CCPA / HIPAA / PCI-DSS / SOX / none]
- **Notification obligations**: [if incident]
- **Handoffs**: Cloak ([reason]), Canon[regulatory] ([reason])

### Recommended Mitigations
- [ ] Canary scope: [% traffic, tenant allowlist]
- [ ] Feature flag kill-switch: [flag name]
- [ ] Pre-deploy verification: [test matrix]
- [ ] Rollback plan: [see rollback-plan output]
- [ ] On-call pre-notification: [team, channel]
- [ ] Status page pre-draft: [yes / no]

### Handoff
- [ ] Builder / Guardian (implementation + PR strategy)
- [ ] Beacon (SLO guard / alert prepare)
- [ ] Triage (on-call brief for SEV1/2)
- [ ] Sentinel (security blast verify)
- [ ] Cloak / Canon[regulatory] (compliance overlay)
- [ ] Launch (rollout gating)

Alert Integration

For changes classified SEV1/SEV2, set pre-deploy alerts in Beacon:

  • Latency SLO: P95 > target × 1.5 → auto-rollback candidate
  • Error rate SLO: 5xx > 1% over 5 min → auto-rollback candidate
  • Business KPI: conversion drop ≥20% over 15 min → pause + investigate

Deliverable Contract

When blast-radius completes, emit:

  • Quantified blast table (all 10 dimensions populated or marked N/A with rationale).
  • SEV tier with explicit dimensional rationale.
  • SLO error-budget burn estimate (minutes, % of monthly).
  • Compliance overlay (regulations, notification obligations, handoffs).
  • Recommended mitigations (canary, feature flag, kill-switch, pre-deploy alerts).
  • Escalation triggers (conditions that would bump tier during rollout).
  • Handoff targets: Beacon (SLO), Triage (SEV1/2 brief), Sentinel (security), Cloak/Canon[regulatory] (data class), Launch (rollout gating).

References

  • Google SRE — Embracing Risk & Error Budgets
  • Amazon 2026 — Mandatory engineering meeting on high-blast-radius AI-assisted changes
  • OWASP 2026 — Top 10 for Agentic Applications (blast radius for agentic systems)
  • Site Reliability Workbook — Incident severity tiers
  • PagerDuty Incident Severity Levels Handbook
  • GDPR Art. 33/34 — Data breach notification
  • HIPAA Breach Notification Rule (45 CFR §§ 164.400-414)

Source: SKILL.md on GitHub

No alerts13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The skill provides comprehensive pre-change impact analysis using dependency mapping and pattern checks. It uses standard developer tools like grep and madge. The primary risk is its ability to process untrusted codebase content which could theoretically contain malicious data, though no specific exploits are currently present in the skill instructions.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    1/5 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at c805268. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/ripple