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 stevengonsalvez/agents-in-a-box --skill sequencegit clone --depth 1 https://github.com/stevengonsalvez/agents-in-a-boxWrote 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/stevengonsalvez/agents-in-a-box/sequence)<a href="https://agentmods.dev/skills/stevengonsalvez/agents-in-a-box/sequence"><img src="https://agentmods.dev/badge/skills/stevengonsalvez/agents-in-a-box/sequence/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/stevengonsalvez/agents-in-a-box/sequence"><img src="https://agentmods.dev/badge/skills/stevengonsalvez/agents-in-a-box/sequence.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00085 | $0.00782 |
| Opus 5 | $0.00043 | $0.00391 |
| Sonnet 5 | $0.00017 | $0.00156 |
| Haiku 4.5 | $0.00009 | $0.00078 |
Grade A, and why
ainb-fleet:sequence 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 9d 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 — 109 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ainb fleet:sequence
Send N prompts in order. Between steps, wait for every target's next assistant-turn-end before firing the next step.
Run
ainb fleet sequence "step1" "step2" "step3" --all --timeout 300
Args
| arg/flag | purpose |
|---|---|
<steps...> |
1 or more prompts, fired in argv order |
--all |
required (v0.1 only supports --all targeting) |
--timeout <seconds> |
per-step ack deadline, default 300 |
How ack-gating works
After each step (except the last):
for each target:
watch ~/.claude/projects/<cwd-slug>/<sid>.jsonl via notify
resolve when next row has:
type == "assistant"
message.stop_reason == "end_turn"
OR timeout fires
This guarantees claude has finished responding to step N before step N+1 lands. The watch runs in parallel across all targets; the next step fires when all targets either acked or timed out.
Common flows
Disconnect → reconnect cycle:
ainb fleet sequence \
"remote-control disconnect" \
"remote-control" \
--all
Clear → resume:
ainb fleet sequence \
"/clear" \
"continue from where you left off" \
--all
Verify-then-act:
ainb fleet sequence \
"/status" \
"if status shows clean, proceed; otherwise abort" \
--all
Caveats
- --all only in v0.1. Filter/cwd selectors not yet wired for sequence. Workaround: use a wrapper script that builds a filter list then runs per-target sequences.
- Ack detection requires JSONL access. If
~/.claude/projects/is unreadable (perms / sandboxing), ack always times out → step N+1 fires after the timeout regardless of actual completion. - Default 300s is generous. For chatty steps (analysis prompts, code
generation), 300s is realistic. For trivial prompts (
/clear), pass--timeout 30to fail fast. - Tail wallclock = sum of per-step wallclocks. A 3-step sequence against a slow model can take minutes. Daemon mode keeps moving; sequence is intentionally synchronous.
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.
- 9d ago First seen · 109 lines · 85 tokens per session scan A 3c00c543cd2f
ainb-fleet:sequence is a skill published in the GitHub repository stevengonsalvez/agents-in-a-box (23 stars, last pushed today), licensed MIT. It adds 85 tokens to every session and 782 once invoked, about $0.0004 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
chinese-git-workflow
A Git workflow guide tailored to Chinese-hosted platforms such as Gitee, Coding.net, and GitLab, including their code-review and build systems.
chinese-commit-conventions
A Chinese-language convention for writing Git commit messages, based on Conventional Commits, a standard format that tools can use to classify changes and build changelogs.
using-superpowers
A set of rules for finding and using assistant skills at the start of every conversation. It says to check for relevant skills before responding, including when asking a clarification question.
platform-api
Queries the Genesys Cloud Platform API schema for endpoint details, permissions, parameters, and response formats. Use when user asks about Genesys Cloud API endpoints (e.g., "/api/v2/assistants/queues"), API documentation, API permissions, schemas, capabilities, or when implementing Genesys Cloud operations.
chinese-code-review
Chinese-language guidance for giving professional, tactful code-review feedback while still calling out real problems.
mcp-builder
A guide to building MCP servers, which let AI assistants call tools or read connected data. It covers server structure, tool design, validation, testing and deployment in TypeScript and Python.