SkillVaultskills Browse all 500 skills

Reliability · Version 1.8.0 · Reviewed 2026-08-02

On-call Assistant

Reduce production risk in live triage and mitigation options with evidence, explicit trade-offs, and a verification plan.

4 method steps 4 documented failure modes 4 diagnostic checks 7 quality gates

Helps a responder triage an active incident: stabilize first, gather evidence, and communicate status clearly.

₹99 one-time

Get this skill archive

What this skill helps you do

  • Live triage
  • Mitigation options
  • Status communication

How On-call Assistant works

You provide

Impact window, telemetry, and dependency state

It inspects

Saturation and blast radius behind live triage

It decides

A mitigation options plan that stabilizes before diagnosing

You verify

Detect, mitigate, and resolve times recorded separately

What it checks first

On-call Assistant helps a responder triage an active incident: stabilize first, gather evidence, and communicate status clearly. Use it when the work involves Live triage, Mitigation options, Status communication.

  1. User-visible impact and error-budget consumption rather than component health.
  2. Saturation signals — queue depth, pool utilization, connection counts — near the onset.
  3. Whether the system recovered on its own, which indicates saturation rather than corruption.
  4. The blast radius and what boundary should have contained it.

Failure modes it recognizes

  • Retry amplification turning a partial failure into a total outage.
  • A shared dependency creating correlated failure across supposedly independent services.
  • Slow resource exhaustion invisible until a hard limit is crossed.
  • A rollback blocked by an incompatible migration.

Answers it will reject

  • Treating the trigger as the root cause, which stops the analysis before the fragility is identified.
  • Adding a runbook step where a boundary would remove the failure mode.
  • Measuring availability as a mean, which hides regional and tenant-level outages.

Decision rules it applies

  • Stabilize user impact before completing diagnosis.
  • Bound every retry with a budget, jitter, and a circuit breaker.
  • Prefer removing a failure mode over detecting it faster.

Evidence it asks for

  • Record time-to-detect, time-to-mitigate, and time-to-resolve separately.
  • Quantify impact in customer terms: failed requests, affected accounts, duration.
  • Verify recovery with the same signal that detected the failure.

The method inside

  1. Reconstruct the symptom timeline and define what healthy behavior would look like for live triage.
  2. Rank hypotheses for mitigation options by evidence, blast radius, and ability to explain every observed symptom.
  3. Run the cheapest discriminating check for status communication; update confidence only when evidence changes.
  4. Separate immediate stabilization, confirmed cause, contributing conditions, and prevention; finish with a reproducible verification.

Deliverables

  • Live triage assessment
  • Mitigation options decision and action plan
  • Status communication verification checklist

Evidence requirements

  • User-visible symptoms and SLO impact
  • Timeline, telemetry, deploys, and dependency state
  • Current mitigations and operational constraints

Quality gates

  • Every material claim traces to supplied evidence or is labeled as a hypothesis.
  • The response follows the declared deliverable contract.
  • No execution, access, measurement, or verification is invented.
  • Secrets and personal data are redacted rather than repeated.
  • The user receives a concrete independent verification step.
  • The relevant failure modes in this domain were considered rather than only the reported symptom.
  • No listed anti-pattern was recommended as a solution.

Example task

Input

Error rate jumped to 12% five minutes ago. Latency is normal. What do I check first?

Expected output

High errors with normal latency suggests fast failures, not saturation. Check recent deploys and config changes in the last 30 minutes before touching capacity...

Boundaries and compatibility

Ideal for

  • Live triage: produce a decision or artifact grounded in supplied evidence.
  • Mitigation options: produce a decision or artifact grounded in supplied evidence.
  • Status communication: produce a decision or artifact grounded in supplied evidence.

Out of scope

  • Replacing incident command authority
  • Calling a trigger the root cause without a causal chain

Agent compatibility

  • GitHub Copilot custom agents
  • Claude Agent Skills / SKILL.md
  • Any instruction-following chat model

Tool policy: Advisory by default. No tools are assumed. If the host provides tools, use read-only evidence gathering unless the user explicitly approves a scoped write or execution action.