cas-worker

A guide for a coding-agent worker operating in a multi-agent factory, where a supervisor assigns and coordinates tasks.

In plain words
What is it for?
Use it to inspect and start assigned tasks, follow acceptance criteria, commit logical changes, add progress or blocker notes, communicate with the supervisor, and close completed work.
Why use it?
It explains how to find assignments, report progress, handle blockers, and preserve work across shared or separate workspaces.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/codingagentsystem/cas/cas-worker
Any agent
npx skills add codingagentsystem/cas --skill cas-worker
Clone the repo
git clone --depth 1 https://github.com/codingagentsystem/cas

Made for: Claude Code, Codex.

Per session 38 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,177 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00038 $0.01177
Opus 5 $0.00019 $0.00589
Sonnet 5 $0.00008 $0.00235
Haiku 4.5 $0.00004 $0.00118

Measured yesterday against content hash 0a6d779f9a2f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cas-worker scanned grade A with 0 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 yesterday.

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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/skills/cas-worker/SKILL.md · 123 lines

How it starts

The opening of the file, as written. The whole thing — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Factory Worker

You execute tasks assigned by the Supervisor. You may be working in an isolated git worktree or sharing the main working directory — check your environment with mcp__cas__coordination action=my_context.

Workflow

  1. Check assignments: mcp__cas__task action=mine
  2. Start a task: mcp__cas__task action=start id=<task-id>
  3. Read task details and understand acceptance criteria before coding: mcp__cas__task action=show id=<task-id>
  4. Implement the solution, committing after each logical unit of work
  5. Report progress: mcp__cas__task action=notes id=<task-id> notes="..." note_type=progress
  6. Close when done: mcp__cas__task action=close id=<task-id>

If close returns verification-required guidance, message the supervisor to handle it.

Blockers

Report immediately — don't spend time stuck:

mcp__cas__task action=notes id=<task-id> notes="Blocked: <reason>" note_type=blocker
mcp__cas__task action=update id=<task-id> status=blocked

Communication

Never use SendMessage. It is blocked in factory mode. Always use CAS coordination:

mcp__cas__coordination action=message target=supervisor message="<response>" summary="<brief summary>"

Use task notes for ongoing updates (note_type=progress|blocker|decision|discovery). The supervisor sees these in the TUI.

Message the supervisor when you complete a task or need help.

Pre-Close Self-Verification (REQUIRED before closing)

Before running mcp__cas__task action=close, verify your own work. The task-verifier will reject you if any of these fail — save yourself the round-trip.

1. No shortcut markers

# Must return zero results in your changed files
rg 'TODO|FIXME|XXX|HACK|unimplemented!|todo!' <changed_files>
rg 'for now|temporarily|placeholder|stub|workaround' <changed_files>

2. All new code is wired up

For every new function, struct, module, route, or handler you created:

# Verify it's actually called/imported somewhere outside its definition
rg 'your_new_function' src/
ast-grep --lang rust -p 'your_new_function($$$)' src/

If zero external references → you built it but didn't wire it in. Fix before closing.

Read the full file on GitHub · 123 lines

Changes

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.

  1. yesterday First seen · 123 lines · 38 tokens per session scan A 0a6d779f9a2f

Subscribe to this mod's changes

cas-worker is a skill published in the GitHub repository codingagentsystem/cas (151 stars, last pushed 5mo ago), licensed MIT. It adds 38 tokens to every session and 1,177 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.