CLI Agent Orchestrator is a tool that coordinates multiple AI coding command-line programs by running them as separate workers in isolated terminal sessions. A supervisor uses it to delegate software tasks to specialist agents in parallel or in sequence while the workers retain their normal command-line capabilities. The catalogue skills operate this orchestration workflow.
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 awslabs/cli-agent-orchestrator --skill cao-memorygit clone --depth 1 https://github.com/awslabs/cli-agent-orchestratorWrote 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/awslabs/cli-agent-orchestrator/cao-memory)<a href="https://agentmods.dev/skills/awslabs/cli-agent-orchestrator/cao-memory"><img src="https://agentmods.dev/badge/skills/awslabs/cli-agent-orchestrator/cao-memory/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/awslabs/cli-agent-orchestrator/cao-memory"><img src="https://agentmods.dev/badge/skills/awslabs/cli-agent-orchestrator/cao-memory.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 | $0.00059 | $0.01267 |
| Opus 5.5 | $0.00024 | $0.00507 |
| Sonnet 5.5 | $0.00012 | $0.00253 |
| Haiku 4.5 | $0.00006 | $0.00127 |
Grade A, and why
cao-memory 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 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.
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 — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CAO Memory
CAO gives every agent a shared, persistent memory. A fact you store in one session is available to a brand-new agent in a later session — even on a different provider. Use it so the user never has to repeat themselves.
These are CAO's cross-provider memory tools (memory_store, memory_recall,
memory_forget), exposed by the CAO MCP server. They are distinct from any
provider-native memory the CLI tool may have.
Scopes and types
Every memory has a scope (where it applies) and a type (what kind of fact it is).
| Scope | Applies to | Use for |
|---|---|---|
project (default) |
This repo / working directory | Conventions, architecture, build rules |
global |
Every project | User identity, durable cross-project preferences |
federated |
Every project on this machine | Reusable, repo-independent lessons worth sharing across all your work (rejects credentials) |
session |
This run only | Short-lived task context |
agent |
This agent role | Role-specific working notes |
Types: project (default), user (who the user is / preferences), feedback
(corrections and how-to-work guidance), reference (pointers to docs, tickets, URLs).
Recall — check memory BEFORE asking the user
At the start of a task, and whenever you're about to ask the user something they may have already told you, search memory first.
memory_recall(query="database widgets endpoint testing")
Omit scope to search all scopes (results follow precedence session → project → global → agent → federated).
Filter with scope= or memory_type= when you know where to look. Recall is for searching
beyond what was auto-injected (see below) — don't re-recall what's already in front of you.
Store — save anything worth remembering, immediately
Store the moment you learn something durable. Don't wait until the end of the session. Store conclusions, not transcript. Keep each memory to 1–2 sentences.
Store when you hit any of these:
- A correction — "No, we use DynamoDB here, not SQL." → store it so no agent makes that mistake again.
- A decided convention — "Every endpoint must have a pytest test before merge."
- A user preference — how they like work done, tools they prefer.
- A non-obvious project constraint — something you couldn't infer from the code.
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 Changed · +6 lines 742ab0f2163b
- 1mo ago First seen · 116 lines · 59 tokens per session scan A 91d1d9c585b5
cao-memory is a skill published in the GitHub repository awslabs/cli-agent-orchestrator (1,480 stars, last pushed today), licensed Apache-2.0. It adds 59 tokens to every session and 1,267 once invoked, about $0.0002 per session on Opus 5.5. 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
tool-selection
Internal guidance for choosing between agent-rack's sync and background execution tools, and when to use per-agent shortcuts. Use whenever delegating a task to a claude/codex/opencode/Antigravity/custom sub-agent through agent-rack.
review
Run a structured, read-only code review through agent-rack's agentreview tool.
session-send
Send follow-up input to a running background agent-rack session.
review-handling
Internal guidance for presenting agent-rack's agentreview output back to the user. Use whenever an agentreview call (foreground or via agentsessionstatus) returns a result.
session-logs
Read the raw stdout/stderr event stream from a background agent-rack session.
session-start
Start a background sub-agent session via agent-rack for a long-running task.