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 skills/cloudchef/atlasclaw-providers/request-decomposition-agentnpx skills add CloudChef/atlasclaw-providers --skill request-decomposition-agentgit clone --depth 1 https://github.com/CloudChef/atlasclaw-providersWrote 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/cloudchef/atlasclaw-providers/request-decomposition-agent)<a href="https://agentmods.dev/skills/cloudchef/atlasclaw-providers/request-decomposition-agent"><img src="https://agentmods.dev/badge/skills/cloudchef/atlasclaw-providers/request-decomposition-agent.svg" alt="Measured on agentmods" 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 | $0.00027 | $0.02338 |
| Opus 5 | $0.00014 | $0.01169 |
| Sonnet 5 | $0.00005 | $0.00468 |
| Haiku 4.5 | $0.00003 | $0.00234 |
Grade A, and why
request-decomposition-agent 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 5d 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 — 289 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Request Decomposition Agent
Orchestration agent for transforming descriptive demands into request candidates.
Purpose
When receiving free-form infrastructure or application requirements:
- Parse and decompose them into executable sub-requests
- Match each sub-request to available catalog services
- Build structured request payloads with resolved and unresolved fields
- Return draft requests for human review
Default behavior is draft-first. In webhook robot admin execution mode, this agent may submit the decomposed child requests only when the webhook selects an authorized robot profile and the input explicitly allows submission.
Trigger Conditions
This skill activates when:
- Input is descriptive text instead of a clean catalog request
agent_identityisagent-request-orchestratorrequest_textis provided
For ordinary chat/runtime routing, this skill should be preferred whenever the user asks for different resource types in one request, or when the user asks for multiple resources with distinct per-item configuration. Do not route here just because the user asks for quantity N of the same resource type with the same parameters.
Robot Admin Execution
For webhook-driven backend execution, AtlasClaw must select an explicit
SmartCMP provider instance and robot/admin credential in the Tool Context. If
that instance is not configured, execution must fail closed rather than
falling back to prod or another instance.
The robot provider instance should use a SmartCMP cmp_tk_* provider token
when available. SmartCMP Provider sends those tokens as
Authorization: Bearer <token> and keep non-cmp_tk_* session tokens on the
existing CloudChef-Authenticate header.
The AtlasClaw webhook user (ATLASCLAW_USER_ID=webhook-*) is the trigger
identity, not the SmartCMP request actor. In webhook robot dispatches without
forwarded SmartCMP user cookies, SmartCMP Provider resolves the SmartCMP
actor from the selected robot credential instead of the synthetic webhook
user id.
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.
- 5d ago First seen · 289 lines · 27 tokens per session scan A 0a6000ae1620
request-decomposition-agent is a skill published in the GitHub repository CloudChef/atlasclaw-providers (15 stars, last pushed 2d ago), licensed Apache-2.0. It adds 27 tokens to every session and 2,338 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 skills, from other repositories
aixyz
Build, run, and deploy an AI agent using the aixyz framework. Use this skill when creating a new agent, adding tools, wiring up A2A/MCP protocols, configuring x402 micropayments, or deploying to Vercel.
cli-for-agent
Design and build CLI tools optimized for AI agent consumption. Use this skill when creating a new CLI command, refactoring an existing CLI for agent compatibility, or reviewing whether a CLI follows agent-friendly design principles.
create-use-agently-cli
Create a new CLI command for the use-agently CLI. Use this skill when implementing a new command, subcommand, or protocol handler in the packages/use-agently package following the CLI design specification.
use-agently
Discover (by searching) to use thousands of AI agents, MCP, tools, data sources, and APIs on the use-agently.com AI Agent Marketplace/Directory. You can get more effective "Web Search", "Web Fetch" or "Social Search" with use-agently. Sending messages via the A2A protocol, calling MCP tools, making HTTP/Web requests…
grizzly_sms
SMS verification and virtual phone numbers via Grizzly SMS API.
quality-assurance
代码审查、质量验收、集成测试、性能测试、交付把关.