Product Decisions This Supports
- Audit trail and compliance readiness: v0.0.10 adds soft-delete support, meaning records removed through the admin panel are flagged as deleted rather than physically purged. This is the difference between an audit log that survives a "cleanup" and one that doesn't — a gating requirement for regulated customers (finance, healthcare, public sector) and for any internal policy that mandates retention of deleted records.
- Data recovery as a user-facing feature: Soft deletes open the door to restore workflows, "recently deleted" views, and undo affordances in the admin UI. Useful for sales/ops teams who currently file support tickets to recover mistakenly deleted records.
- Roadmap sequencing: If "compliance-grade audit logging" or "admin undo/restore" sits on the roadmap, this release removes a blocker. If it doesn't, this is a nice-to-have that doesn't change positioning.
- Build vs. buy: Adding soft deletes across an existing model layer is deceptively expensive — migrations, relationship scoping, unique-index collisions, restore-path testing. A package that handles this at the audit layer is a credible argument for buy over build, provided the underlying schema assumptions match ours.
- Release cadence signal: The jump from v0.0.9 to v0.0.10 in a 0.0.x line tells you this is pre-1.0 software with no stability guarantee. Treat it as an evaluation candidate, not a production dependency, until it earns a 1.0.
When to Consider This Package
Adopt for evaluation when:
- You run Filament admin panels and need an audit/activity layer without writing one from scratch.
- Retaining deleted records is a real requirement — regulatory, contractual, or operational (fraud investigation, dispute resolution, customer support forensics).
- Your team is small enough that maintaining a bespoke audit-log implementation is a poor use of engineering time.
- You can pin the version and absorb occasional breaking changes in a 0.0.x line.
Look elsewhere when:
- You need a stable, long-term-supported dependency with a published deprecation policy — 0.0.x is explicitly unstable, and minor releases can move under you.
- Your data model has heavy soft-delete interactions (polymorphic relations, cascading restores, unique constraints that must exclude trashed rows). Validate against your own schema before committing; the release note is one line, and the depth of the implementation isn't documented.
- Your compliance regime requires certified, independently audited tooling rather than an open-source plugin with a single recent contributor.
- You only need a thin activity log — Laravel's first-party activity-logging options may be sufficient and cheaper to own.
Immediate diligence before any pilot: confirm whether soft deletes are captured by default or opt-in, whether restore events are also logged, and how existing hard-deleted or already-trashed records are handled on upgrade.
How to Pitch It (Stakeholders)
For executives: This release closes a real gap. Until now, records deleted through our admin panel were gone — the audit trail described a deletion but couldn't reproduce the deleted record. With soft-delete support, deletions become reversible and auditable, which is what compliance reviewers and incident investigations actually ask for. It's a small release note for a meaningful capability. The caveat: the package is pre-1.0 and maintained by a very small team, so this is a "pilot and monitor" recommendation, not an "adopt org-wide" one.
For engineering: v0.0.10 adds soft-delete handling to the Filament Blackbox audit layer. Worth a spike to verify it matches our model conventions — specifically how it interacts with existing SoftDeletes traits, whether restore is logged as a distinct event, and the migration cost on our largest tables. The release notes are thin, so read the diff before trusting the changelog. Recommend pinning the version, running it against a staging copy of our most relationally complex model, and revisiting once soft-delete coverage is confirmed end-to-end.
Recommendation: Pilot, don't standardize. This feature is worth the evaluation effort; the version number is not yet worth a production commitment.