Prepare release verification from a real change
Create a release checklist tied to changed behavior, dependencies, and rollback evidence.
Tamper-evident audit trail for agent workspaces. Every workspace
---
name: openclaw-ledger
description: Tamper-evident audit trail for agent workspaces. Every workspace
change is recorded in a hash-chained log — if anyone alters an entry, the
chain breaks and you know.
---
# OpenClaw Ledger
Tamper-evident audit trail for agent workspaces. Every workspace change is recorded in a hash-chained log — if anyone alters an entry, the chain breaks and you know.
## The Problem
Agents modify files, execute commands, install skills — and leave no verifiable record. If something goes wrong, you can't trace what happened. If logs exist, nothing proves they haven't been altered after the fact.
## Commands
### Initialize
Create the ledger and snapshot current workspace state.
```bash
python3 {baseDir}/scripts/ledger.py init --workspace /path/to/workspace
```
### Record Changes
Snapshot current state and log all changes since last record.
```bash
python3 {baseDir}/scripts/ledger.py record --workspace /path/to/workspace
python3 {baseDir}/scripts/ledger.py record -m "Installed new skill" --workspace /path/to/workspace
```
### Verify Chain
Verify the hash chain is intact — no entries tampered with.
```bash
python3 {baseDir}/scripts/ledger.py verify --workspace /path/to/workspace
```
### View Log
Show recent ledger entries.
```bash
python3 {baseDir}/scripts/ledger.py log --workspace /path/to/workspace
python3 {baseDir}/scripts/ledger.py log -n 20 --workspace /path/to/workspace
```
### Quick Status
```bash
python3 {baseDir}/scripts/ledger.py status --workspace /path/to/workspace
```
## How It Works
Each entry contains:
- Timestamp
- SHA-256 hash of the previous entry
- Event type and data (file changes, snapshots)
If any entry is modified, inserted, or deleted, the hash chain breaks and `verify` detects it.
## Exit Codes
- `0` — Clean / chain intact
- `1` — No ledger or minor issues
- `2` — Chain tampered / corrupt entries
## No External Dependencies
Python standard library only. No pip install. No network calls. Everything runs locally.
## Cross-Platform
Works with OpenClaw, Claude Code, Cursor, and any tool using the Agent Skills specification.
---
name: openclaw-ledger
description: Tamper-evident audit trail for agent workspaces. Every workspace
change is recorded in a hash-chained log — if anyone alters an entry, the
chain breaks and you know.
---
# OpenClaw Ledger
Tamper-evident audit trail for agent workspaces. Every workspace change is recorded in a hash-chained log — if anyone alters an entry, the chain breaks and you know.
## The Problem
Agents modify files, execute commands, install skills — and leave no verifiable record. If something goes wrong, you can't trace what happened. If logs exist, nothing proves they haven't been altered after the fact.
## Commands
### Initialize
Create the ledger and snapshot current workspace state.
```bash
python3 {baseDir}/scripts/ledger.py init --workspace /path/to/workspace
```
### Record Changes
Snapshot current state and log all changes since last record.
```bash
python3 {baseDir}/scripts/ledger.py record --workspace /path/to/workspace
python3 {baseDir}/scripts/ledger.py record -m "Installed new skill" --workspace /path/to/workspace
```
### Verify Chain
Verify the hash chain is intact — no entries tampered with.
```bash
python3 {baseDir}/scripts/ledger.py verify --workspace /path/to/workspace
```
### View Log
Show recent ledger entries.
```bash
python3 {baseDir}/scripts/ledger.py log --workspace /path/to/workspace
python3 {baseDir}/scripts/ledger.py log -n 20 --workspace /path/to/workspace
```
### Quick Status
```bash
python3 {baseDir}/scripts/ledger.py status --workspace /path/to/workspace
```
## How It Works
Each entry contains:
- Timestamp
- SHA-256 hash of the previous entry
- Event type and data (file changes, snapshots)
If any entry is modified, inserted, or deleted, the hash chain breaks and `verify` detects it.
## Exit Codes
- `0` — Clean / chain intact
- `1` — No ledger or minor issues
- `2` — Chain tampered / corrupt entries
## No External Dependencies
Python standard library only. No pip install. No network calls. Everything runs locally.
## Cross-Platform
Works with OpenClaw, Claude Code, Cursor, and any tool using the Agent Skills specification.
Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Includes source scripts; review them before use. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.