Redaction preflight
Known credential, token, email, user-path, source-control, IP, and opaque-value patterns receive stable placeholders without changing line count. Manual review remains mandatory.
This working preview first offers a narrow, line-preserving redaction preflight for known credential and identity patterns. It then normalizes volatile Unity CI output, suppresses lines already present in a last-green run, fingerprints the first new error chain, and applies a small visible ruleset only when the Unity version and build target are comparable.
Known credential, token, email, user-path, source-control, IP, and opaque-value patterns receive stable placeholders without changing line count. Manual review remains mandatory.
Known volatile timestamps, local paths, source coordinates, durations, and long identifiers are normalized before the first new error chain is fingerprinted.
Lines already present in the successful reference are removed from consideration. The preview reports only the first new error or exception chain it can observe.
A small deterministic ruleset can choose one diagnostic branch. It cannot establish root cause, guarantee safe redaction, blame a change, or replace the complete build context.
Unity Build Automation exposes build paths and source-control metadata to jobs, while GitHub warns that automatic secret redaction is not guaranteed and recommends manual log review. The preflight therefore reports its narrow rules and never labels output “safe.”
The private pilot adds an agreed redaction policy, runner metadata, source-line evidence, versioned rules, report output, and an evaluator-owned success test. XeSoft does not request credentials or production secrets.
Discuss a private pilot →