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/marcusquinn/aidevops/fleetgit clone --depth 1 https://github.com/marcusquinn/aidevopsWhat 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 | $0.00018 | $0.00481 |
| Opus 5 | $0.00009 | $0.00241 |
| Sonnet 5 | $0.00004 | $0.00096 |
| Haiku 4.5 | $0.00002 | $0.00048 |
Grade A, and why
fleet 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 — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Use this command guide for aidevops fleet flows that coordinate Vault devices,
encrypted sync, secure messages, and remote-control policy. Arguments: $ARGUMENTS
Safety contract
- Fleet transports are untrusted delivery mechanisms. Git, object storage, secure messaging, SSH, Tailscale/VPN, and VPS hosts may carry encrypted bundles plus signed metadata only.
- Device trust is explicit and revocable. Do not trust a device because it can reach the same account, network, remote, bucket, or message channel.
- Passphrases, recovery material, private keys, raw tokens, and protected plaintext never belong in chat, arguments, environment variables, logs, comments, PR bodies, screenshots, or fixtures.
Fleet flow map
- Onboard device: create local keys, exchange public fingerprints, verify via human/local authenticated channel, sign registry update from a trusted device, then rewrap only allowed collection keys.
- Sync collection: export/import encrypted append-only records; reject unsigned, expired, replayed, rolled-back, tampered, or revoked-device records.
- Secure message: send encrypted envelopes and signed acknowledgements; plaintext subjects, local paths, client names, and private basenames must not appear in transport-visible paths or payloads.
- Remote lock: allow authorized signed protective commands and require a signed acknowledgement with status and audit head.
- Unlock-request: queue a local operator decision and scope any approval to named collections, labels, and duration.
- True remote unlock: default-disabled; only proceed when a future audited policy explicitly proves passphrases never travel through unsafe channels.
Worker metadata
When dispatching fleet work, brief workers with needs_vault,
needs_collections, needs_device, needs_remote_unlock,
data_classification, and runtime_policy. Use collection names and device
classes only; do not include secret values, private paths, client names, or
plaintext message subjects.
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 · 48 lines · 18 tokens per session scan A 689d212734d5
fleet is a command published in the GitHub repository marcusquinn/aidevops (391 stars, last pushed 2d ago), licensed MIT. It adds 18 tokens to every session and 481 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 2026-08-30.
Other commands, from other repositories
gsd:new-project
Initialize a new project with deep context gathering and PROJECT.md.
gsd:new-milestone
Start a new milestone cycle — update PROJECT.md and route to requirements.
gsd:help
Show available GSD commands and usage guide.
gsd:plan-phase
Create detailed execution plan for a phase (PLAN.md) with verification loop.
gsd:execute-phase
Execute all plans in a phase with wave-based parallelization.
gsd:audit-milestone
Audit milestone completion against original intent before archiving.