workflow-guard-rails
Wrap multi-step agent workflows with pre-execution checks, side-effect queues, result validation, retry budgets, checkpointing, audit logs, and failure-rule accumulation. Prevents false successes, duplicate sends, unrecoverable crashes, and silent drift in LLM production systems. Use it when a workf
as observed 2026-10-04T09:46:17.065Z- Identifier
workflow-guard-rails- Source
- ClawHub
- Version observed
- 1.0.6
- Source repository
- not published
- Repository observation
- No source repository listed
- First observed here
- 2026-08-28T22:06:28.212Z
- Observations recorded
- 4
- Installs (reported upstream)
- 1
- Weekly downloads (upstream)
- 613
- Declared license
- MIT-0
Observation history
2026-10-04T09:46:17.065Z
Fields that differed: changelog latestVersion summary
| Field | Before | After |
|---|---|---|
changelog |
"Add a companion-file section to both READMEs pointing at the public error-playbook mirror, so a fresh clone knows what guard #7 writes to and where to fetch it. Also removes three | "Structural fixes per scanner findings: two-stage rule accumulation (same-turn capture into ERROR-PLAYBOOK.staging.md; promotion to the authoritative playbook only on recurrence or |
latestVersion |
"1.0.5" | "1.0.6" |
summary |
"Wrap multi-step agent workflows with pre-execution checks, side-effect queues, result validation, retry budgets, checkpointing, audit logs, and failure-rule accumulation. Prevents | "Wrap multi-step agent workflows with pre-execution checks, side-effect queues, result validation, retry budgets, checkpointing, audit logs, and failure-rule accumulation. Prevents |
2026-09-20T22:20:20.880Z
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"Guard #7 reversed: the agent appends confirmed failures to the error playbook in the same turn, human review is rollback not approval, so unattended runs still accumulate rules. A | "Add a companion-file section to both READMEs pointing at the public error-playbook mirror, so a fresh clone knows what guard #7 writes to and where to fetch it. Also removes three |
latestVersion |
"1.0.4" | "1.0.5" |
2026-09-18T20:20:38.833Z
Fields that differed: license changelog description latestVersion summary
| Field | Before | After |
|---|---|---|
license |
null | "MIT-0" |
changelog |
"Align name with slug; expand retrieval keywords (EN+ZH), read_when triggers, and tags to widen entry points" | "Guard #7 reversed: the agent appends confirmed failures to the error playbook in the same turn, human review is rollback not approval, so unattended runs still accumulate rules. A |
description |
"Wrap multi-step agent workflows with pre-execution checks, side-effect queues, result validation, retry budgets, checkpointing, audit logs, and failure-rule accumulation. Prevents | null |
latestVersion |
"1.0.3" | "1.0.4" |
summary |
"Wrap multi-step agent workflows with pre-execution checks, side-effect queues, result validation, retry budgets, checkpointing, audit logs, and failure-rule accumulation. Prevents | "Wrap multi-step agent workflows with pre-execution checks, side-effect queues, result validation, retry budgets, checkpointing, audit logs, and failure-rule accumulation. Prevents |
Correction
If you maintain this extension and believe anything above is inaccurate, request a correction. Corrections are published, and disputed entries are marked as disputed while under review.