agent-transcript

To add agent transcripts to PR bodies without exposing sensitive logs, use a local transcript-sanitization script that strips private data and inserts a redacted section. This offline process requires no network access and enforces disclosure controls for safe provenance.

Yes, you can preview a redacted transcript before inserting it into a GitHub issue. The script enforces a preview-before-insertion flow, generating a local HTML preview to verify sensitive data is stripped prior to updating the issue body.

No, the agent-transcript script does not require network access to sanitize logs. It operates entirely locally, using best-effort, locally-scoped processing to redact agent transcripts and append them to PR or issue bodies offline.

To provide provenance for agent-generated PRs while keeping logs private, run a local script that appends a sanitized transcript section to PR bodies. This ensures reviewers get necessary context while maintaining offline data privacy.

You should use local transcript sanitization for OpenClaw workflows when reviewers need context from agent-driven PRs or issues but session logs contain sensitive data. It offers a deterministic, privacy-first approach to append redacted provenance.