gastown: Command for Claude Code

.claude/commands/patrol.md

patrol is a command for Claude Code from gastownhall/gastown. It costs 19 tokens per session (1,233 once invoked), scanned A, original, MIT.

A command that runs one monitoring cycle for an agent-support role such as witness, deacon, or refinery. It checks messages and cleanup work, then creates or resumes the role's assigned work item.

In plain words
What is it for?
Use it to run a witness, deacon, or refinery patrol, process pending messages, handle cleanup items, and resume an existing patrol when one is already active.
Why use it?
Background agent work can leave behind unfinished tasks, dirty workspaces, or messages needing attention. A patrol cycle gives the selected role a defined routine for finding and handling them.

Command for Claude Code

Written for Claude Code: allowed-tools in frontmatter.

This is gastownhall/gastown's own configuration. It tells Claude Code how to work on gastown itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything gastown configures →

About the project

Gas Town is a workspace manager for coordinating multiple AI coding agents across projects while preserving their work state in git-backed records. Developers use it to organize agents, tasks, identities, mailboxes, and handoffs across repositories. Its catalogue add-ons support operating and coordinating Gas Town workflows.

gastownhall/gastown · 18,711 stars · on GitHub

Reuse

Borrowing it

Nothing to install: this file belongs to gastownhall/gastown. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/gastownhall/gastown/main/.claude/commands/patrol.md
Clone the repo
git clone --depth 1 https://github.com/gastownhall/gastown

Made for: Claude Code.

Wrote 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.

agentmods badge for patrol

README.md
[![agentmods](https://agentmods.dev/badge/commands/gastownhall/gastown/patrol/github.svg)](https://agentmods.dev/commands/gastownhall/gastown/patrol)
Your own site
<a href="https://agentmods.dev/commands/gastownhall/gastown/patrol"><img src="https://agentmods.dev/badge/commands/gastownhall/gastown/patrol/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.

agentmods 80×15 button for patrol

Your own site · 80×15
<a href="https://agentmods.dev/commands/gastownhall/gastown/patrol"><img src="https://agentmods.dev/badge/commands/gastownhall/gastown/patrol.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 19 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,233 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00019 $0.01233
Opus 5.5 $0.00008 $0.00493
Sonnet 5.5 $0.00004 $0.00247
Haiku 4.5 $0.00002 $0.00123

Measured 2mo ago against content hash 1f6948be8cdf, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

patrol 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 2mo 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.

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/commands/patrol.md · 146 lines

How it starts

The opening of the file, as written. The whole thing — 146 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Patrol

Run one patrol cycle for the specified role. Patrol is a continuous monitoring loop — each invocation executes one cycle of the patrol formula.

Arguments: $ARGUMENTS If no role specified, detect from current GT_ROLE environment variable.

Role Detection

echo $GT_ROLE

Map to patrol type:

  • */witness → witness patrol (mol-witness-patrol)
  • */deacon or */deacon/* → deacon patrol (mol-deacon-patrol)
  • */refinery → refinery patrol (mol-refinery-patrol)
  • Explicit argument overrides detection

Patrol Entry Point

gt patrol new --role <role>

This creates a hooked wisp with steps from the patrol formula. If a patrol is already running (wisp exists on hook), resume it instead.

Witness Patrol Steps

The witness is the per-rig polecat supervisor. Execute in order:

1. inbox-check

gt mail inbox

Process any pending messages: POLECAT_DONE, MERGED, HELP, escalations. Read each with gt mail read <id> and take appropriate action.

2. process-cleanups

Check for cleanup wisps (dirty state from dead polecats):

bd list --status=open --label=cleanup --json

Process each: verify git state, clean worktrees, close cleanup wisps.

3. check-refinery

gt peek gastown/refinery

Verify refinery is alive and processing the merge queue. If stuck, nudge: gt nudge gastown/refinery "Health check — are you processing?"

4. survey-workers

Check all active polecats in the rig:

gt peek gastown/polecats

For each active polecat:

  • Check if session is alive (has recent activity)
  • Check if work is progressing (commits, bead updates)
  • Detect zombies: session dead but agent_state says working
  • Detect stale spawns: spawning > 10 minutes

Nudge idle polecats:

gt nudge gastown/polecats/<name> "Progress check — what's your status?"

5. check-timer-gates

bd list --label=gate:timer --status=open --json

Evaluate any timer-based gates that may have elapsed.

6. check-swarm

Check for convoy/swarm completion across coordinated work:

bd list --label=convoy --status=open --json

Read the full file on GitHub · 146 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. 2mo ago First seen · 146 lines · 19 tokens per session scan A 1f6948be8cdf

Subscribe to this mod's changes

patrol is a command published in the GitHub repository gastownhall/gastown (18,711 stars, last pushed 9d ago), licensed MIT. It adds 19 tokens to every session and 1,233 once invoked, about $0.0001 per session on Opus 5.5. 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.

Related

patrol compared with commands from other repositories

patrol next to 6 commands from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
patrolgastownhall/gastownCommand19k9d agoMITA19Claude Code
checklistgithub/spec-kitCommand148kyesterdayMITA13Claude Code
clarifygithub/spec-kitCommand148kyesterdayMITA27Claude Code
specifygithub/spec-kitCommand148kyesterdayMITA13Claude Code
convergegithub/spec-kitCommand148kyesterdayMITA35Claude Code
implementgithub/spec-kitCommand148kyesterdayMITA14Claude Code
speckit.bug.fixgithub/spec-kitCommand148kyesterdayMITA12Claude Code