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 skills add MSApps-Mobile/claude-plugins --skill session-backupgit clone --depth 1 https://github.com/MSApps-Mobile/claude-pluginsWrote 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/msapps-mobile/claude-plugins/session-backup)<a href="https://agentmods.dev/skills/msapps-mobile/claude-plugins/session-backup"><img src="https://agentmods.dev/badge/skills/msapps-mobile/claude-plugins/session-backup/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.
<a href="https://agentmods.dev/skills/msapps-mobile/claude-plugins/session-backup"><img src="https://agentmods.dev/badge/skills/msapps-mobile/claude-plugins/session-backup.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00036 | $0.02456 |
| Opus 5.5 | $0.00014 | $0.00982 |
| Sonnet 5.5 | $0.00007 | $0.00491 |
| Haiku 4.5 | $0.00004 | $0.00246 |
Grade E, and why
daily-cowork-backup scanned grade E with 4 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 1mo 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.
Enumerates the file system for secretshighData exfiltration
Searching home directories for .env, .ssh, .aws or credential files is reconnaissance for credential theft.
candidates = glob.glob('/sessions/*/mnt/Claude/credentials/cowork-gdrive-config.json') Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf /tmp/cowork-backup /tmp/upload_payload.json Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
> **Why the two-step POST→GET?** Google Apps Script returns a 302 redirect to `script.googleusercontent.com`. curl's `-L` flag converts POST→GET for 302s, but the redirect endpoint still returns HTML. The fix: POST witho Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
r1 = subprocess.run( How it starts
The opening of the file, as written. The whole thing — 255 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Back up all Cowork-related folders to Google Drive.
Session path note: Cowork VM session names change every run. Do NOT hardcode a session name like
loving-gracious-allen. Instead, discover the active session dynamically (Step 1).
Step 1: Discover session path and read GDrive config
import json, os, glob, sys
# Discover the active session that has the credentials file
candidates = glob.glob('/sessions/*/mnt/Claude/credentials/cowork-gdrive-config.json')
if not candidates:
print('STATUS:NOT_FOUND')
sys.exit(1)
cfg_path = candidates[0]
session_mnt = cfg_path.replace('/Claude/credentials/cowork-gdrive-config.json', '') # e.g. /sessions/abc-xyz/mnt
with open(cfg_path) as f:
cfg = json.load(f)
print('STATUS:OK')
print(f'SESSION_MNT:{session_mnt}')
print(f'URL:{cfg["url"]}')
print(f'APIKEY:{cfg["apiKey"]}')
If STATUS is NOT_FOUND, stop and report:
Config file missing. On your Mac, run once:
cp ~/.cowork-gdrive-config.json ~/Documents/Claude/credentials/cowork-gdrive-config.json
Step 2: Flush debug screenshots
Delete temporary debug screenshots from the Claude docs folder (top level only).
Why Desktop Commander? The VM filesystem mount is read-only for deletes — use Desktop Commander to run the delete natively on the Mac.
Use mcp__Desktop_Commander__start_process with:
command: find ~/Documents/Claude -maxdepth 1 -type f \( -name '*.png' -o -name '*.jpg' -o -name '*.jpeg' -o -name '*.gif' \) -delete && echo 'Screenshots flushed'
timeout_ms: 15000
(Failure is non-fatal — continue anyway.)
Step 3: Build backup zips into /tmp/cowork-backup/
All zips must exclude: */node_modules/*, */.git/*, */.next/*, *.pyc, *.png, *.jpg, *.jpeg, *.gif, */package-lock.json
Use SESSION_MNT from Step 1 for all paths.
3a: Skills
mkdir -p /tmp/cowork-backup
cd '{SESSION_MNT}/.claude' && zip -r /tmp/cowork-backup/skills-backup.zip skills/ \
-x '*/node_modules/*' -x '*/.git/*' -x '*/.next/*' -x '*.pyc' \
-x '*.png' -x '*.jpg' -x '*.jpeg' -x '*.gif' -x '*/package-lock.json' \
2>&1 | tail -2 && ls -lh /tmp/cowork-backup/skills-backup.zip
What ships with it
1 file 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.
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.
- 1mo ago First seen · 255 lines · 36 tokens per session scan E 5f87022d07e4
daily-cowork-backup is a skill published in the GitHub repository MSApps-Mobile/claude-plugins (9 stars, last pushed today), licensed MIT. It adds 36 tokens to every session and 2,456 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it E with 4 findings (enumerates the file system for secrets, recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-13.
Other skills, from other repositories
rented-server-setup
Runs an open model in the cloud two ways, and makes the user choose with real numbers. Route A is OVHcloud AI Endpoints, an EU-owned per-token API with zero idle cost, for spiky usage and EU jurisdiction. Route B rents one single-tenant RunPod GPU pod in a region the user picks, serving the model on vLLM behind a…
allow-team
Puts a running Open WebUI or Odysseus on a shareable public URL through an ngrok tunnel, so a teammate can sign in and use the local models without installing anything. Use when the user says "share Open WebUI with my team", "share Odysseus with my team", "let my colleague use my local AI", "expose Open WebUI", "share…
azure-mcp-observability
The read-only observability half of the azure-mcp connector — the monitor, resourcehealth, applens, and advisor namespaces: Azure Monitor metrics, Log Analytics KQL, alert-rule state, platform health states, AppLens detectors, and Advisor recommendation categories, plus the degraded-resource investigation order.
Axcient Clients
Axcient x360Recover clients: the healthstatus model, clientcode, the three-way protected-system counter breakdown (appliancebased, d2c, cloudarchive), and minting direct-to-cloud agent enrollment tokens.
Axcient Jobs
Axcient x360Recover backup jobs: the BRC vs Replibit job-type split, per-job protection thresholds, and job run history with its starttimebegin pagination floor.
Axcient Vaults
Axcient x360Recover vaults: Private vs Cloud vault types, storage capacity, and the connectivity-loss alert threshold — including axcientsetvaultthreshold, the one tool in this plugin that changes alerting configuration rather than just reading it.