feat: retry with backoff on changelog step failures #500

Problem

The changelog step can fail transiently (network issues, rate limits, temporary API unavailability). When it does, the entire workflow fails with no retry, forcing a manual re-run.

Example failure:

• Run: https://github.com/LerianStudio/github-actions-shared-workflows/actions/runs/28115689968/job/83254695338 (attempt ci(deps): bump actions/setup-node from 4 to 5 #1)

Proposed Solution

Add a retry-with-backoff mechanism to the changelog step:

• Minimum 3 retry attempts before marking the step as failed
• Exponential backoff between attempts (e.g. 10s, 30s, 90s or similar)
• Log each attempt clearly so failures are traceable
• Only fail the workflow after all retries are exhausted

Implementation Options

• Use a retry action (e.g. nick-fields/retry or similar) wrapping the changelog step
• Or implement a shell loop with sleep and attempt counter inline in the step

Impact

Reduces false-positive workflow failures caused by transient changelog errors, removes the need for manual re-runs on flaky steps.