salesforce-incident-runbook

v2026.09.24

Run evidence-led Salesforce integration incident response from detection through containment, recovery, reconciliation, and prevention. Use when handling service or data incidents. Trigger with "respond to Salesforce incident".

GitHub
安装命令
npx skhub add jeremylongshore/salesforce-incident-runbook
Markdown
SKILL.md

Salesforce Integration Incident Response

Overview

Restore business invariants safely by separating Salesforce platform state, customer org state, application behavior, data correctness, and event delivery.

Prerequisites

  • Incident commander, severity model, affected environments and workloads, communications, and evidence channel
  • Current deployment, org identity, app and principal, API, limits, jobs, events, data, and reconciliation contracts
  • Pre-approved containment, failover, deferral, rollback, credential revocation, and vendor escalation procedures

Tool Discipline

Use Read, Glob, and Grep to inspect approved repository and evidence files, WebFetch to re-check current first-party Salesforce documentation, and Write or Edit only for secretless plans, fixtures, configuration, and redacted receipts.

Current Contract

Salesforce Status reports platform incidents, while org-specific authorization, limits, metadata, automation, jobs, events, and data can fail independently. A green platform status does not prove an integration or business invariant is healthy.

Authentication

Use read-only incident access first and separate break-glass access under the customer process. Never paste tokens, sessions, private keys, raw records, or unrestricted logs into tickets or chat.

Instructions

  1. Declare severity, commander, time window, affected orgs and workloads, business invariant, data risk, and communication cadence.
  2. Freeze deployments and unsafe retries; preserve request, job, event, deployment, and reconciliation identifiers.
  3. Check Salesforce Status, org identity, authorization, API support, limits, metadata, automation, async jobs, event flow, and application telemetry.
  4. Classify platform, org, identity, permission, schema, capacity, code, dependency, data, or consumer failure and bound impact.
  5. Select the least-destructive approved containment such as pausing intake, deferring work, isolating a consumer, rollback, or credential revocation.
  6. Recover in a bounded cohort, reconcile records and events by stable keys, and monitor technical and business stop signals.
  7. Close only after invariants, backlog, security, and data exposure are resolved; assign root-cause and prevention work.

Approval Boundaries

Do not rotate credentials, broaden access, disable automation, replay events, retry writes, purge queues, roll back, or fail over without incident and system owners.

Output

Return severity, timeline, evidence, classification, containment, approved actions, recovery IDs, reconciliation, exposure assessment, communications, and prevention owners.

Error Handling

ConditionResponse
Salesforce Status is greenContinue org and application diagnosis; do not downgrade from platform status alone.
Prior write outcome is unknownBlock replay and reconcile by stable business key and request evidence.
Containment can lose business workUse durable deferral or quarantine and document the recovery obligation.

Example

A redacted completion receipt might look like this:

incident=SF-311; severity=2; class=event-consumer; intake=paused; recovery=bounded; backlog=0; reconcile=exact

Resources

Next Steps

Run the workflow first in the lowest-risk authorized org and preserve its redacted receipt. Schedule a review against the next Salesforce seasonal release and the customer change calendar.

发现
标签

此技能尚未发布标签。

版本
最新版本元数据

版本

v2026.09.24

发布时间

2026年9月24日

分类

未分类

许可证

MIT

源路径

skills/.curated/salesforce-incident-runbook

默认分支

main

最新提交

e5a6c3b

Tree SHA

c2dc8e8