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/motlin/claude-code-plugins/restorenpx skills add motlin/claude-code-plugins --skill restoregit clone --depth 1 https://github.com/motlin/claude-code-pluginsWhat 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 | $0.00070 | $0.00891 |
| Opus 5 | $0.00035 | $0.00445 |
| Sonnet 5 | $0.00014 | $0.00178 |
| Haiku 4.5 | $0.00007 | $0.00089 |
Grade A, and why
restore 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 2d 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.
What it actually says
Restore Herdr Workspaces
After a reboot, rebuild the session from the snapshot written by /herdr-reboot:snapshot: every
captured workspace with its label, every tab with its label, the panes inside each tab split at the
captured direction and ratio, then the resume command fired into each new pane.
Preview the plan first:
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/restore.py .llm/resume-after-reboot-state.json
The preview prints the tree it will build — workspaces, tabs, split right 0.6 lines, and the
pane that each slot lands in. Show it to the user. Once they confirm, fire it:
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/restore.py .llm/resume-after-reboot-state.json --go
Use --limit N to rebuild only the first N workspaces and --skip 3,7 to leave individual slots
out entirely; a split left holding one pane collapses, and a tab left holding none is not created.
Agent launches are spaced 0.4s apart (--delay) so a couple dozen boots do not all land at once.
The state file is JSON, schema resume-after-reboot/v2, and is herdr-shaped. A tmux-reboot
snapshot (resume-after-reboot/v1) is flat, with no workspaces, tabs, or splits to rebuild, so
this restore rejects it rather than half-restoring it.
What the user needs to know before firing:
commandpanes split on theirrestore_defaultfield. Dev servers (restore_default: false) are skipped: they often survive the reboot that killed the agents, and re-running one just fails withEADDRINUSEagainst the server still holding the port. Their pane is still created, so the user can start them by hand. Pass--commandsto fire them anyway, only after confirming the old processes are gone.- Read-only viewers (
restore_default: true—git log,less FILE,htop,tig) fire without--commands. Nothing of theirs survives a reboot, so theEADDRINUSEreasoning does not apply. They re-run from the top: scroll position, search, and selection are not restored. - Nothing is ever typed into a pane that is not an idle shell.
herdr pane runtypes into whatever the pane holds, so firing into a resumed agent would submit the resume command to it as a prompt. Every pane is fired only when confirmed to be sitting at a shell prompt with no live agent; anything else is skipped with a printed reason. Re-running the script is therefore safe. - A live workspace with the same label and working directory is adopted rather than duplicated. Its captured tabs land there as brand new tabs, never reusing its live panes.
- Focus comes back last: each workspace's active tab, then the workspace that held focus. Focus inside a tab is restored as the tree is built, by creating the focused pane focused.
- Verify codex panes whose note says
herdr reported no session. Those ids come from matching the working directory against rollout files on disk, so the pairing is a best guess. If a resumed session opens the wrong conversation, runcodex resumein that pane and use the interactive picker. claude-rcpanes are Remote Control servers, fired like any other agent pane. Their command reattaches with--continue, which errors out if nothing was recorded for that directory in roughly the last 4 hours — expect that on any reboot you did not come straight back from. The pane is left at a prompt; start a fresh server there by hand when you want one.
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.
- 2d ago First seen · 59 lines · 70 tokens per session scan A 60eae2a6bc06
restore is a skill published in the GitHub repository motlin/claude-code-plugins (15 stars, last pushed 3d ago), licensed Apache-2.0. It adds 70 tokens to every session and 891 once invoked, about $0.0003 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.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…