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.
npx agentmods add skills/forgeyclap/claude-forge/forge-nightshiftnpx skills add ForgeyClap/claude-forge --skill forge-nightshiftgit clone --depth 1 https://github.com/ForgeyClap/claude-forgeWrote 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/forgeyclap/claude-forge/forge-nightshift)<a href="https://agentmods.dev/skills/forgeyclap/claude-forge/forge-nightshift"><img src="https://agentmods.dev/badge/skills/forgeyclap/claude-forge/forge-nightshift.svg" alt="Measured on agentmods" 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 | $0.00043 | $0.01985 |
| Opus 5 | $0.00022 | $0.00992 |
| Sonnet 5 | $0.00009 | $0.00397 |
| Haiku 4.5 | $0.00004 | $0.00198 |
Grade A, and why
forge-nightshift 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 4d 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.
How it starts
The opening of the file, as written. The whole thing — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Forge playbook — Nightshift (overnight self-resuming builder)
Do not duplicate existing tools — defer to: forge-manifest.cjs (ARM/RECONCILE per-WP manifest),
forge-swarm-resume.cjs (re-dispatch plan for unfinished WPs), forge-checkpoint.cjs (idempotency for any
side-effecting WP), forge-actiongate.cjs (hard-gates — deploy/push/spend/DNS/etc.), usage-guard.cjs
(the real, already-built session/usage-limit watchdog + auto-resume signal), and forge-briefing.cjs
(piece J5 — the real, tested morning-briefing generator this skill's output depends on). This file is
doctrine/orchestration only: it explains how those pieces compose into an overnight run, it does not
reimplement any of them.
What Nightshift actually is (and is not)
Nightshift is not a new autonomous scheduler and not a background daemon that decides on its own to start work. It is a documented pattern for composing four already-real, already-tested pieces so a large swarm can survive a session/usage-limit interruption and resume honestly the next time a session runs:
- Plan the work queue up front and ARM it (
forge-manifest.cjs::arm()) so every work package's exact scope (narrowed_prompt,deps) is persisted BEFORE dispatch — never re-derived from memory later. - Dispatch normally (Agent-tool subagents, per this project's real-agents-only governance). Each
WP's real completion/failure is logged as an ordinary event (
wp_completed/check_passedorwp_failed/check_failed) — Nightshift adds no new event-logging mechanism of its own. - Survive the interruption. If the session ends (session-limit pause, usage-limit pause, or the owner
simply closes the laptop), nothing is lost: the manifest +
events.jsonlalready on disk are the complete record. The NEXT session that runs callsforge-swarm-resume.cjs::resume({run_id})to get back exactly the unfinished WPs (with their original scoping) — never a fabricated "pick up where I think I left off". - Report honestly. Before (or instead of) resuming, run
forge-briefing.cjs --run <id>to generate the MORNING BRIEFING — a markdown of what ran, what's blocked, and what only the owner can decide (retry/reassign/drop a failure, resume/reprioritize/drop an unfinished WP), derived purely from the run's own logged manifest + events. This is a real, hermetically-tested tool (forge-bin/forge-briefing.cjs+forge-briefing.test.cjs), not a scaffold.
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.
- 4d ago First seen · 100 lines · 43 tokens per session scan A 5be2aac7d528
forge-nightshift is a skill published in the GitHub repository ForgeyClap/claude-forge (2 stars, last pushed 1mo ago), licensed MIT. It adds 43 tokens to every session and 1,985 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 2026-08-31.
Other skills, from other repositories
frame-macos-notification
拟真 macOS 通知 banner + app icon + 标题正文, 适合 video overlay / 产品发布预告.
deck-course-module
暖纸背景 + Playfair, 左侧学习目标常驻, 含 MCQ 自测页.
deck-pitch
10 页融资 deck, 白底 + 蓝紫渐变 hero, traction 柱状, $X.XM ask.
argent-tv-interact
Control and inspect TV apps via argent — Apple TV (tvOS), Android TV (leanback), and Amazon Fire TV (Vega). Boot the target, read focus, navigate with the D-pad remote, type, screenshot, and on Vega debug the JS runtime (evaluate, console logs, network inspector). Use when a task targets a TV (runtimeKind "tv", or…
review-offered-task
Review a task that has been offered to you and decide whether to accept or reject it.
aidd-dev:08:for-sure
Iterative agent loop that tracks attempts and retries until a success condition is met. Use when the user says "for sure", "make sure", "keep trying until", "loop until done", "don't stop until", or needs guaranteed completion of a task with explicit success criteria.