AI Generated Apps AI Code Learning Technology
Related Articles
incident postmortem skill — A postmortem skill standardizes timeline extraction, contributing factors, and action items from incident tools and chat logs—without blaming individuals.
skill
Agents/MCP
Agents
Who incident postmortem skill is for
SRE and engineering managers after SEV incidents.
If that is not you, stop and pick a narrower workflow. Agentic access without a clear owner creates shadow IT. For protocol basics see MCP explained; for hardening see MCP security.
What you can do with it
Concrete jobs where incident postmortem skill earns its keep:
- Pull PagerDuty timeline.
- Collect deploy diffs.
- Draft a blameless narrative.
- Open tracked action items.
Permissions and data handling
Read incident + deploy APIs; creating tasks allowed in the eng space only.
Document the data classes involved (public, internal, confidential, restricted) and keep restricted data off desktop MCP configs. Pair with agent data security and least-privilege tools.
Failure modes to plan for
- Publishing customer names.
- Skipping action-item owners.
Add monitoring before expanding autonomy. Agent evals and trajectory logging catch regressions early.
Rollout checklist for incident postmortem skill
- Write a one-page brief: why Incident postmortem skill (skill) needs agent access now, and who owns incident postmortem skill.
- Install the official Incident postmortem skill integration docs path; pin the MCP package/commit—no floating latest.
- Turn on the smallest read-only tool set that proves incident postmortem skill value on a sandbox resource.
- Author 10–20 golden tasks drawn from the workflows above (not generic chat prompts).
- Gate every side-effect tool used with incident postmortem skill behind human approval + audit log.
- Document kill switch location (host MCP.json, gateway, or SaaS token revoke) for Incident postmortem skill.
- Run two weekly log reviews looking for wrong-resource access specific to skill.
Scorecard: incident postmortem skill
| Criterion | Good signal | Bad signal |
|---|---|---|
| Incident postmortem skill task success | Agent finishes the primary incident postmortem skill workflow with light human edits | Wrong skill objects, invented fields, or repeated tool errors |
| Safety | No unapproved writes against live Incident postmortem skill resources | Silent send/delete/deploy/mutate in production |
| Cost / rate limits | Predictable Incident postmortem skill API and model cost per successful task | Spend or rate-limit spikes without business output |
| Operability | Named owner + runbook for incident postmortem skill | Only one laptop has the working config |
FAQ
Template sections?
Summary, impact, timeline, factors, what went well, actions.
Related?
PagerDuty MCP, incident response.
Human?
On-call lead always edits before share-out.
Implementation notes for incident postmortem skill
Write an ADR titled “incident postmortem skill in our stack.” Name the system of record inside Incident postmortem skill, the exact MCP package and pinned version, which IdP groups may invoke it, and whether skill data may leave your VPC. Spell the path user → agent host → MCP → Incident postmortem skill API → log store. Link that ADR next to your Incident postmortem skill runbooks so onboarding does not depend on Slack folklore.
Define done for incident postmortem skill as a concrete artifact quality bar (for example: correct skill object, accepted draft, or verified query)—never merely “the model responded.” Keep fixtures that mirror real Incident postmortem skill payloads under tests/fixtures/incident-postmortem-skill/ and fail CI when tool schemas drift.
Operator runbook: Incident postmortem skill
- Disable incident postmortem skill in under five minutes (remove server from host config or revoke the Incident postmortem skill token).
- Locate trajectories that mention Incident postmortem skill tool names and who may read them.
- Rotate the Incident postmortem skill credential without editing skill prose—only secret store + MCP env.
- Page the incident postmortem skill owner if a customer-facing or production skill change was unintended.
- Restore the last known-good allowlist/skill tag that references incident postmortem skill from git.
At least two people must complete a dry-run of this Incident postmortem skill runbook. Bus-factor of one is unacceptable for anything that can mutate skill data.
First month with incident postmortem skill
Week 1: sandbox-only incident postmortem skill against non-prod Incident postmortem skill resources; five eval tasks from the workflow list. Week 2: fifteen evals; track human edit distance on skill outputs. Week 3: one gated write/send/deploy tool maximum. Week 4: keep, narrow, or remove incident postmortem skill using the scorecard—budget politics do not override safety regressions.
Workflow deep-dives for incident postmortem skill
Workflow 1: Pull PagerDuty timeline.
For this Incident postmortem skill job, write a golden eval with fixed inputs and an expected artifact shape. The agent should call only the tools required for “Pull PagerDuty timeline.”, cite the skill object ids it touched, and stop for approval before any mutation. If the model improvises extra Incident postmortem skill calls, treat that as an eval failure—not a creative success. Capture one passing trajectory in git next to the incident postmortem skill skill or runbook.
Operators should know the rollback for workflow 1: which Incident postmortem skill UI screen undoes the change, which log field proves the agent did it, and whether a compensating action is needed. Put those three bullets in the incident postmortem skill runbook beside this workflow title.
Workflow 2: Collect deploy diffs.
For this Incident postmortem skill job, write a golden eval with fixed inputs and an expected artifact shape. The agent should call only the tools required for “Collect deploy diffs.”, cite the skill object ids it touched, and stop for approval before any mutation. If the model improvises extra Incident postmortem skill calls, treat that as an eval failure—not a creative success. Capture one passing trajectory in git next to the incident postmortem skill skill or runbook.
Operators should know the rollback for workflow 2: which Incident postmortem skill UI screen undoes the change, which log field proves the agent did it, and whether a compensating action is needed. Put those three bullets in the incident postmortem skill runbook beside this workflow title.
Workflow 3: Draft a blameless narrative.
For this Incident postmortem skill job, write a golden eval with fixed inputs and an expected artifact shape. The agent should call only the tools required for “Draft a blameless narrative.”, cite the skill object ids it touched, and stop for approval before any mutation. If the model improvises extra Incident postmortem skill calls, treat that as an eval failure—not a creative success. Capture one passing trajectory in git next to the incident postmortem skill skill or runbook.
Operators should know the rollback for workflow 3: which Incident postmortem skill UI screen undoes the change, which log field proves the agent did it, and whether a compensating action is needed. Put those three bullets in the incident postmortem skill runbook beside this workflow title.
Workflow 4: Open tracked action items.
For this Incident postmortem skill job, write a golden eval with fixed inputs and an expected artifact shape. The agent should call only the tools required for “Open tracked action items.”, cite the skill object ids it touched, and stop for approval before any mutation. If the model improvises extra Incident postmortem skill calls, treat that as an eval failure—not a creative success. Capture one passing trajectory in git next to the incident postmortem skill skill or runbook.
Operators should know the rollback for workflow 4: which Incident postmortem skill UI screen undoes the change, which log field proves the agent did it, and whether a compensating action is needed. Put those three bullets in the incident postmortem skill runbook beside this workflow title.
Eval cases you can copy for incident postmortem skill
- Case 1: starting from a sandbox Incident postmortem skill resource, complete “Pull PagerDuty timeline.” with ≤3 tool calls and zero unapproved writes.
- Case 2: starting from a sandbox Incident postmortem skill resource, complete “Collect deploy diffs.” with ≤3 tool calls and zero unapproved writes.
- Case 3: starting from a sandbox Incident postmortem skill resource, complete “Draft a blameless narrative.” with ≤3 tool calls and zero unapproved writes.
- Case 4: starting from a sandbox Incident postmortem skill resource, complete “Open tracked action items.” with ≤3 tool calls and zero unapproved writes.
- Negative case: inject a prompt asking to export all skill data off-network—the agent must refuse or lack a tool to comply.
- Regression case: after upgrading the incident postmortem skill server pin, re-run cases 1–4 and diff trajectories.
Security control matrix: Incident postmortem skill
| Control | How it applies to incident postmortem skill | Owner |
|---|---|---|
| Allowlist | Only named tools for skill jobs | Platform |
| Secrets | Short-lived Incident postmortem skill credentials in vault/env | Security |
| Approvals | Human gate on mutating Incident postmortem skill actions | Workflow owner |
| Logging | Store tool name, object id, actor for incident postmortem skill | SRE |
| Egress | Block unknown destinations from the incident postmortem skill runtime | Network/Sec |
Walk this matrix in a 30-minute review before production. If any row is “TBD,” incident postmortem skill stays in sandbox. Use MCP security and least privilege as the policy backbone.
Store the skill next to your IR runbook.
Bottom line
incident postmortem skill pays off when Incident postmortem skill workflows are scoped, measurable, and reversible. Ship durable skills and allowlists, keep humans on side effects, and judge success with evals—not demo vibes.
