Borrowing it
Nothing to install: this file belongs to ric03uec/clawrium. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/ric03uec/clawrium/main/.claude/skills/clawctl-lmwork/SKILL.mdgit clone --depth 1 https://github.com/ric03uec/clawriumWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ric03uec/clawrium/clawctl-lmwork)<a href="https://agentmods.dev/skills/ric03uec/clawrium/clawctl-lmwork"><img src="https://agentmods.dev/badge/skills/ric03uec/clawrium/clawctl-lmwork/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ric03uec/clawrium/clawctl-lmwork"><img src="https://agentmods.dev/badge/skills/ric03uec/clawrium/clawctl-lmwork.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00050 | $0.08699 |
| Opus 5.5 | $0.00020 | $0.03480 |
| Sonnet 5.5 | $0.00010 | $0.01740 |
| Haiku 4.5 | $0.00005 | $0.00870 |
Grade B, and why
clawctl-lmwork scanned grade B with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
chmod 600 "$TMP" Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
The orchestrator **never calls the local stack directly.** No `curl` to How it starts
The opening of the file, as written. The whole thing — 705 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Local-Model Work Dispatch
Keep the local inference stack busy on clawrium work it can actually finish, and drive that work to a reviewed PR without babysitting.
This skill is clawrium-only. The selection criteria and the staleness checks below encode this repo's conventions; they do not transfer.
Roles
Three actors. Do not invent a fourth.
| Role | Is | Where |
|---|---|---|
| orchestrator | the session running this skill | your current window |
| lmworker | pi interactive, Qwen3.6-27B via vllm-inx |
window lmworker |
| lmjudge | claude --dangerously-skip-permissions interactive |
window lmjudge |
The orchestrator never calls the local stack directly. No curl to
LiteLLM, no API calls. It always spins up a pi harness and types into
it, exactly as a human would in an interactive session. That is the
entire point of driving this through tmux.
The orchestrator also does not touch source code, run tests, or
review diffs. lmworker writes; lmjudge reviews; lmworker calls ATX.
This mirrors the orchestration contract in
itx-execute.
Modes
scan— harvest + staleness-check + classify, labelagent-ready, dispatch nothing.<issue-number>— run the full loop for one issue.status— report in-flight sessions and reconcile labels (see below).- no args —
scan, then dispatch the top T1 candidate.
Preflight
Refuse to start unless all of these hold:
ssh inx 'docker ps --format "{{.Names}}"' | grep -q vllm-qwen3.6-27b-fp8 # stack up
command -v pi && command -v tmux && command -v atx # tooling
atx server status | grep -q 'Running: true' # atx daemon
atx project status # project registered
Cap 2 issues in flight — vLLM runs num_seq=3, and lmjudge is
Anthropic-side so it does not contend for the local stack.
The loop
| # | Step | Owner |
|---|---|---|
| 0 | Preflight; count in-flight sessions | orchestrator |
| 1 | Harvest candidates | orchestrator |
| 2 | Staleness check → DISPATCH / RESCOPE / CLOSE-REC / TRAP |
orchestrator |
| 3 | Classify T1 / T2 / park; auto-approve only xs/s |
orchestrator |
| 4 | Create worktree + session + two windows; stamp start time | orchestrator |
| 5 | Write <worktree>/.itx/<N>/lmwork-brief.md |
orchestrator |
| 6 | Claim issue (in-progress); send brief to lmworker window |
orchestrator |
| 7 | Implement, commit locally | lmworker |
| 8 | Wait for idle; send review request to lmjudge window |
orchestrator |
| 9 | Review; verdict SATISFIED or REVISE |
lmjudge |
| 10 | REVISE → relay findings into the live pi session; back to 7. Max 3 rounds |
orchestrator |
| 11 | SATISFIED → send the ATX command to lmworker |
orchestrator |
| 12 | Run ATX, fix blockers. Max 3 iterations | lmworker |
| 13 | Open the PR; flip issue to in-review |
lmjudge / orchestrator |
| 14 | Comment the outcome on the issue | orchestrator |
| 15 | Append to ~/.local/state/clawrium/lmwork-ledger.jsonl |
orchestrator |
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 705 lines · 50 tokens per session scan B 1f6e69c9365b
clawctl-lmwork is a skill published in the GitHub repository ric03uec/clawrium (67 stars, last pushed 13d ago), licensed Apache-2.0. It adds 50 tokens to every session and 8,699 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it B with 2 findings (asks for root, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-08.
Other skills, from other repositories
gh
GitHub CLI skill for interacting with GitHub via the gh command line tool. Use when Bub needs to (1) Create, view, or manage GitHub repositories, (2) Work with issues and pull requests, (3) Create and manage releases, (4) Run and monitor GitHub Actions workflows, (5) Create and manage gists, (6) Install and manage…
python
Use when the task is Python itself, in any framework or none: PEP 695 generics, mypy --strict typing, dataclass/Protocol/TypedDict/Enum choices, asyncio.TaskGroup, stdlib idioms, src/ layout + pyproject.toml with uv, ruff+mypy+pytest gate. NOT a FastAPI/ASGI service (that is fastapi), NOT a deep pytest suite (that is…
pypi-release
This skill should be used when releasing tunacode-cli to PyPI. It keeps the existing local release checks, then hands the actual PyPI upload to a GitHub Actions workflow that uses the repository's PYPIAPITOKEN secret.
python-format
Formats Python files with ruff format and reports which files changed.
spawnpoint
This skill should be used when an agent needs to create, list, extend, or remove multi-repo git worktree workspaces using the spawnpoint CLI (binary spawnpoint, shell wrapper sp). Use it when a task involves spinning up isolated worktrees for a feature branch across one or more repos, finding existing workspaces…
agentic-preflight
Use when shipping a branch: reviewing, documenting, linting, testing, and pushing work through Agentic Preflight. Also use when a push is blocked by the agentic-preflight pre-push hook or when the user says agentic-preflight:uninstall to remove this tool from the current project.