A setup for running a project-building process across repeated work cycles, with one coordinator assigning milestones to worker agents. It stores progress in files and uses Git branches, commits, and pull requests to track changes.

Use it to coordinate an overnight or multi-day build across many milestones, including application code, documentation, tests, and screenshots in one repository.

It keeps a long autonomous build from losing its place between cycles. It also provides a recorded roadmap, shared state, and checks for deciding whether each milestone is complete or blocked.

long-run — Skill for Claude Code by reidworks-io/reidworks-skills

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.

Any agent

npx skills add reidworks-io/reidworks-skills --skill long-run

Clone the repo

git clone --depth 1 https://github.com/reidworks-io/reidworks-skills

Made for: Claude Code.

Or install reidworks-core, the plugin that ships this one along with the rest of its 3 skills.

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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/reidworks-io/reidworks-skills/long-run/github.svg)](https://agentmods.dev/skills/reidworks-io/reidworks-skills/long-run)

Your own site

<a href="https://agentmods.dev/skills/reidworks-io/reidworks-skills/long-run"><img src="https://agentmods.dev/badge/skills/reidworks-io/reidworks-skills/long-run/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/reidworks-io/reidworks-skills/long-run"><img src="https://agentmods.dev/badge/skills/reidworks-io/reidworks-skills/long-run.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 166 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 3,766 The whole file, excluding the scripts and references it only reads on demand.

Security scan B 1 finding. 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.

Model	Per session	Once invoked
Fable 5.1	$0.00166	$0.03766
Opus 5.5	$0.00066	$0.01506
Sonnet 5.5	$0.00033	$0.00753
Haiku 4.5	$0.00017	$0.00377

Measured 2mo ago against content hash 79b9943c0e4f, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade B, and why

long-run scanned grade B with 1 finding 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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

test -f .claude/settings.local.json && grep -q '"deny"' .claude/settings.local.json && echo "✓ deny list" || echo "✗ no deny list"

plugins/reidworks-core/skills/long-run/SKILL.md · 305 lines

How it starts

The opening of the file, as written. The whole thing — 305 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Long Run — autonomous multi-agent build orchestrator

Distilled from a working pattern (weekend SDK + monorepo build, 14+ milestones, file-based state, /loop-driven). Build the harness; the user owns the kickoff.

Operating model

One orchestrator, many workers. A /loop cycle re-fires every N minutes. Each cycle the orchestrator picks the next eligible milestone, spawns a worker via the Agent tool, runs gates, merges (or files a block), updates state, exits the turn.
State on disk, not in conversation. Everything that matters across cycles lives in .agents/STATE.md, .agents/DIGEST.md, .agents/BLOCKED.md. The orchestrator is stateless between cycles and reconstructs from these files every time.
Git is the source of truth. Workers branch, commit, open PRs, squash-merge. The diff is the artifact. No deployments, no force pushes.
Roadmap is gospel. Never invent milestones; only execute them.
One repo, one prototype. App, docs, marketing, tests, screenshots all colocated so a single PR can update related surfaces atomically.

Hard prerequisites

Do not proceed to scaffolding until ALL are true. If any is missing, engage the user.

Git repository. test -d .git.
Detailed roadmap at .agents/ROADMAP.md (or equivalent) with explicit per-milestone:
- id (e.g. M0, P2.1)
- dependencies (other milestone ids)
- contract (exact allowed-paths whitelist)
- DoD (verifiable bullets, not "make it good")
- model + subagent type
Source-of-truth docs in repo for the surfaces in scope:
- DESIGN.md if UI work is in scope
- VOICE.md if user-facing copy is in scope
- architecture doc if SDK/library work is in scope
Quality gates wired up. A single command that runs typecheck + lint + test + build. Workers must pass it before merge.
Deny-list in .claude/settings.local.json blocking deploy commands and force-push.
.agents/STOP convention as kill switch.

Read the full file on GitHub · 305 lines

Files

What ships with it

7 files 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.

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.

2mo ago First seen · 305 lines · 166 tokens per session scan B 79b9943c0e4f

Subscribe to this mod's changes

long-run is a skill published in the GitHub repository reidworks-io/reidworks-skills (1 stars, last pushed 3mo ago), licensed MIT. It adds 166 tokens to every session and 3,766 once invoked, about $0.0007 per session on Opus 5.5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Building agents Claude Code Planning ai-tools claude claude-skills

long-run next to 6 skills from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
long-runreidworks-io/reidworks-skills	Skill	1	3mo ago	MIT	B	166	Claude Code
skill-builderyusufkaraaslan/Skill_Seekers	Skill	15k	1mo ago	MIT	A	38	Claude Code, Codex
rudder-data-graphsrudderlabs/rudder-agent-skills	Skill	19	yesterday	MIT	A	40	Claude Code
rudder-typer-workflowrudderlabs/rudder-agent-skills	Skill	19	yesterday	MIT	A	43	Claude Code
rudder-instrumentation-planningrudderlabs/rudder-agent-skills	Skill	19	yesterday	MIT	A	32	Claude Code
rudder-tracking-reviewrudderlabs/rudder-agent-skills	Skill	19	yesterday	MIT	A	66	Claude Code
rudder-ai-events-verifyrudderlabs/rudder-agent-skills	Skill	19	yesterday	MIT	A	66	Claude Code

long-run