BambooHR Privacy-Safe Debug Receipt
Overview
Create evidence sufficient to reproduce and route a BambooHR failure without packaging employee records, credentials, tokens, raw request/response bodies, webhook payloads, or environment dumps.
Prerequisites
- The target repository or integration path and the requested operator outcome.
- The tenant, identity, and data scope only when approved live work is in scope.
- The current evidence register plus customer-specific permissions and agreements.
Current Contract
The official Python SDK surfaces a request ID from x-request-id,
x-bamboohr-request-id, or request-id and includes a secure log filter for
sensitive headers and URL parameters. These controls do not make arbitrary
application logs safe; verify the emitted artifact field by field.
Authentication
Diagnostics may record only auth mode, credential owner alias, token/key age, and last rotation time. Never include Authorization headers, API keys, client secrets, refresh/access tokens, cookie values, or webhook private keys.
Instructions
- Define incident ID, tenant alias, UTC window, operation, expected result, actual status, impact, and evidence recipient.
- Locate only bounded logs for the affected request IDs. Do not grep or export an entire environment, home directory, database, or log bucket.
- Produce a manifest containing application version/commit, SDK or adapter version, endpoint template without IDs/query values, auth mode, timeout, configured retry count, status, latency, request IDs, and safe aggregate counts. Do not retain the response body.
- Replace tenant, employee, file, benefit, applicant, and webhook identifiers with stable incident-local aliases. Remove bodies and free-text errors that may echo BambooHR's response content.
- Scan the proposed receipt for secret formats, email addresses, government IDs, dates of birth, addresses, compensation, health/benefit data, and URLs containing query values. Review manually after automated scanning.
- Write a checksum and explicit inclusion/exclusion list. Keep the artifact local until the data owner approves its recipient and retention.
- Prefer a directory or JSON/Markdown receipt. Do not create a broad tarball by default; archive only the reviewed manifest if the support channel requires it.
Tool Discipline
Use Read, Glob, and Grep only against the approved paths and time window. Use Write/Edit to create the minimal receipt and redaction tests. Do not use shell archive or network tools under this skill.
Approval Boundaries
Require approval before reading production logs, writing a diagnostic artifact, including any pseudonymized employee fact, or sending evidence to BambooHR or a third party. Creation does not authorize transmission.
Output
Return receipt path, checksum, incident window, request IDs, included safe fields, excluded sensitive classes, scanner/manual-review result, approved recipient, retention deadline, and transmission status.
Error Handling
- Suspected secret or PII in output: stop, quarantine locally, and rebuild from the allowlist; do not attempt line-by-line salvage for transmission.
- Missing request ID: correlate on a narrow timestamp/operation window and label the result lower confidence.
- Recipient asks for raw payload: escalate to the data owner/security process.
Examples
- "Zip all BambooHR logs" becomes a manifest-first, allowlisted receipt.
- "Support needs the failing employee JSON" substitutes request ID, status, and schema.
Resources
Read official evidence before collecting evidence.