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 2026-11-19T03:34:19.311Z- Identifier
workflow-guard-rails- Source
- ClawHub
- Version observed
- 1.0.10
- Source repository
- not published
- Repository observation
- No source repository listed
- First observed here
- 2026-08-28T22:06:28.212Z
- Observations recorded
- 8
- Installs (reported upstream)
- 1
- Weekly downloads (upstream)
- 1,177
- Declared license
- MIT-0
Observation history
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"Maintenance: 七守卫表与 Hard Rule 7 要求'同 turn'捕获失败,Operating procedure 却写'same run',跨多 turn 工作流会放宽为延迟捕获,统一为同 turn。" | "Review fix: workflow-guard-rails — 修复扫描 finding(行为更严谨/权限声明/脱敏)。" |
latestVersion |
"1.0.9" | "1.0.10" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"Add negative invariants to guard-5 validation (no placeholders, stale markers, duplicate items)" | "Maintenance: 七守卫表与 Hard Rule 7 要求'同 turn'捕获失败,Operating procedure 却写'same run',跨多 turn 工作流会放宽为延迟捕获,统一为同 turn。" |
latestVersion |
"1.0.8" | "1.0.9" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"Add checkpoint trap: validate checkpoints with guard-5 assertions on resume; a corrupt checkpoint defends the wrong output" | "Add negative invariants to guard-5 validation (no placeholders, stale markers, duplicate items)" |
latestVersion |
"1.0.7" | "1.0.8" |
Fields that differed: changelog latestVersion
| Field | Before | After |
|---|---|---|
changelog |
"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 | "Add checkpoint trap: validate checkpoints with guard-5 assertions on resume; a corrupt checkpoint defends the wrong output" |
latestVersion |
"1.0.6" | "1.0.7" |
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 |
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" |
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.