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 commands/callmetechie/fleet-manager/first-rungit clone --depth 1 https://github.com/CallMeTechie/fleet-managerWhat 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 | $0.00014 | $0.00808 |
| Opus 5 | $0.00007 | $0.00404 |
| Sonnet 5 | $0.00003 | $0.00162 |
| Haiku 4.5 | $0.00001 | $0.00081 |
Grade A, and why
first-run scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
First Run
set -euo pipefail
source "${CLAUDE_PLUGIN_ROOT:-plugin}/commands/_fleet-lib.sh"
mkdir -p "$FM_SERVERS_DIR"
if [ -n "$(list_server_names)" ]; then
echo "Inventory already has servers. Use /add-server to add another, or /list-servers to view them."
exit 0
fi
echo "No servers yet — starting intake."
If the inventory is empty, run the intake dialogue below yourself, here in the
main thread. The fleet-intake subagent has no AskUserQuestion tool — subagents
cannot reach the user — so it is dispatched only once every answer is in hand.
1. Intake dialogue (you, in the main thread)
Ask via AskUserQuestion, one question at a time. Validate each answer before
moving on and re-ask on a violation:
| Ask for | Constraint |
|---|---|
| Server name | ^[a-zA-Z0-9_-]+$ |
| Host (LAN IP, hostname or WAN domain) | ^[a-zA-Z0-9.-]+$ |
| SSH port | 1–65535, default 22 |
| SSH user | ^[a-zA-Z0-9_.-]+$ |
| Description | free text, optional, multiword allowed |
| Authorized scopes (multi-select) | any of system_monitoring, file_operations, package_management, service_management, docker_compose, file_transfer |
Offer scopes as an explicit multi-select. Do not preselect them all — each one widens what later commands may do without asking again.
2. Dispatch the agent
Use Claude's native Task tool (NOT a bash command — do not write Task(...)
in a script) to invoke the subagent named fleet-intake (Claude Code resolves
plugin agents by their frontmatter name), passing every collected value:
Onboard this server into the fleet-manager workspace. It is non-interactive: all inputs are below, and you have no way to ask for more. name= host= port= user= description= scopes= set_active=yes Write the Connection section, ensure the plugin keypair exists, cold-test the connection, then run discovery, apply exactly these scopes, and regenerate inventory.md. If the cold test fails, return NEEDS_KEY_DEPLOY with the exact ssh-copy-id line and stop.
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 · 70 lines · 14 tokens per session scan A 981be40c73f4
first-run is a command published in the GitHub repository CallMeTechie/fleet-manager (1 stars, last pushed 29d ago), licensed MIT. It adds 14 tokens to every session and 808 once invoked, about $0.0001 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 commands, from other repositories
dns-provider
Apply the CLI runtime and safety contract before executing a command. Runtime --help is authoritative.
domains
Apply the CLI runtime and safety contract before executing a command. Runtime --help is authoritative.
cloud
Apply the CLI runtime and safety contract before executing a command. Runtime --help is authoritative.
instances
Apply the CLI runtime and safety contract before executing a command. Runtime --help is authoritative.
deploy-static
Apply the CLI runtime and safety contract before executing a command. Runtime --help is authoritative.
cluster
Apply the CLI runtime and safety contract before executing a command. Runtime --help is authoritative.