ClawHub

OpenClaw Model Usage

v1.2.1

Inspect local OpenClaw model usage directly from session logs. Use when asked for the current model, recent model usage, usage breakdown by model, token tota...

⭐ 0· 117·0 current·0 all-time
byRana Salal Ali@ranasalalali
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name/description match the code and instructions: the package includes a Python CLI and script that parse ~/.openclaw/agents/*/sessions/*.jsonl and sessions.json to produce usage summaries and an optional local HTML dashboard. Nothing requested (env vars, binaries, config paths) is unrelated to this purpose.
✓
Instruction Scope
SKILL.md and the CLI are scoped to reading local session JSONL and sessions.json files, producing summaries and a self-contained HTML file. The instructions explicitly direct the agent to use a local root (default ~/.openclaw/agents) and warn not to use test fixtures for production. There are no instructions to collect or transmit data to external endpoints.
✓
Install Mechanism
No install spec is provided (instruction-only at registry level) and the project bundles Python scripts and a CLI implementation. This is low-risk: nothing is fetched from external URLs and no archive extraction or remote installs are present in the metadata. Running the tool requires a local Python runtime (pyproject requires >=3.12).
✓
Credentials
The skill requires no environment variables, credentials, or special config paths beyond reading local OpenClaw logs. The requested file access (user home ~/.openclaw/agents) is coherent with the described functionality. No unrelated secrets or external service keys are requested.
✓
Persistence & Privilege
The skill is not forced-always (always:false) and does not modify other skills or global agent configuration. It writes output only when asked (e.g., --out for dashboard) and otherwise performs read-only analysis of local logs.
Assessment
This skill appears to do what it says: parse local OpenClaw session JSONL files and build local summaries and an HTML dashboard. Before installing/running: (1) confirm you have Python 3.12+ available (the package requires it), (2) be aware it will read files under ~/.openclaw/agents by default (so don't run it if those logs are sensitive or you don't want local analysis), (3) the dashboard is written locally (--out path) and is self-contained (no external scripts), and (4) no credentials or network calls are requested by the code—if you need extra assurance, inspect the bundled scripts (scripts/model_usage.py and src/openclaw_model_usage/cli.py) or run the tool in a sandboxed environment. The smoke tests exercise subprocesses for local invocation but those are test-only and not network-related.

Like a lobster shell, security has layers — review code before you run it.

latestvk9755yd6z6f0mr5hjf6j0a8bq184321p

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments