llm-cli-gateway: Skill for Claude Code

.agents/skills/async-job-orchestration/SKILL.md

async-job-orchestration is a skill for Claude Code from verivus-oss/llm-cli-gateway. It costs 75 tokens per session (7,427 once invoked), scanned A, original, MIT.

A guide for running long-lasting language-model jobs without blocking the caller, across several supported coding assistants. It covers job storage, duplicate prevention, caching, and provider-specific launch rules.

In plain words
What is it for?
Dispatching asynchronous jobs, choosing safe provider defaults, handling cache state, and applying the correct settings for Claude, Codex, Gemini, Grok, Mistral, Devin, and Cursor.
Why use it?
It helps jobs survive gateway restarts when durable storage is configured and keeps each assistant's command-line options and approval behavior correct.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: mentions Claude Code; installed under .agents/ (shared by several agents); mentions Codex.

This is verivus-oss/llm-cli-gateway's own configuration. It tells Claude Code how to work on llm-cli-gateway itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything llm-cli-gateway configures →

Part of the llm-gateway plugin — 24 skills, 4 commands, 2 MCP servers shipped together

Reuse

Borrowing it

Nothing to install: this file belongs to verivus-oss/llm-cli-gateway. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/verivus-oss/llm-cli-gateway/main/.agents/skills/async-job-orchestration/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/verivus-oss/llm-cli-gateway

Made for: Claude Code.

Or install llm-gateway, the plugin that ships this one along with the rest of its 24 skills, 4 commands, 2 MCP servers.

Wrote 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.

agentmods badge for async-job-orchestration

README.md
[![agentmods](https://agentmods.dev/badge/skills/verivus-oss/llm-cli-gateway/async-job-orchestration/github.svg)](https://agentmods.dev/skills/verivus-oss/llm-cli-gateway/async-job-orchestration)
Your own site
<a href="https://agentmods.dev/skills/verivus-oss/llm-cli-gateway/async-job-orchestration"><img src="https://agentmods.dev/badge/skills/verivus-oss/llm-cli-gateway/async-job-orchestration/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.

agentmods 80×15 button for async-job-orchestration

Your own site · 80×15
<a href="https://agentmods.dev/skills/verivus-oss/llm-cli-gateway/async-job-orchestration"><img src="https://agentmods.dev/badge/skills/verivus-oss/llm-cli-gateway/async-job-orchestration.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 75 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 7,427 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector warn 7 Sept 2026
SkillSpector: 8 findings, up to high

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

  • high Rogue Agent · line 542
    Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.
    Fix: Prevent the skill from modifying its own code, SKILL.md, or configuration files. Treat skill files as read-only at runtime.
  • medium Prompt Injection · line 289
    Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
    Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
  • medium Prompt Injection · line 292
    Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
    Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
  • medium Prompt Injection · line 293
    Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
    Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
  • medium Prompt Injection · line 294
    Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
    Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
  • medium Prompt Injection · line 295
    Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
    Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
  • medium Prompt Injection · line 296
    Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
    Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
  • medium Rogue Agent · line 404
    Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
    Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
How audits are shown
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00075 $0.07427
Opus 5 $0.00037 $0.03714
Sonnet 5 $0.00015 $0.01485
Haiku 4.5 $0.00007 $0.00743

Measured 10d ago against content hash 2399047b5d7e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

async-job-orchestration 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 10d 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.

.agents/skills/async-job-orchestration/SKILL.md · 556 lines

How it starts

The opening of the file, as written. The whole thing — 556 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Async Job Orchestration

Async execution is available for Claude, Codex, Gemini, Grok, Mistral, Devin, and Cursor when the gateway registers async tools. SQLite and PostgreSQL stores preserve jobs across gateway restarts; acknowledged memory storage lasts only for the process lifetime; persistence.backend = "none" registers no async or job tools. See Durability & Dedup.

Mistral Vibe: the gateway always emits --agent <mode> and defaults to accept-edits for programmatic callers. Mistral uses approvalStrategy:"legacy": mcp_managed is rejected before launch and approvalPolicy has no effect. Current Vibe defaults session logging on; doctor --json flags an explicit [session_logging] enabled = false before session-continuity use.

Dispatch Defaults

Apply these on every dispatch unless the caller has explicitly overridden a rule in the current turn:

  1. Omit model: let the gateway use its configured default per CLI. Nominating a model risks deprecated IDs (o3, o3-pro, gpt-4o, …) and capability mismatches.
  2. Managed approval is Claude-only. Use approvalStrategy:"mcp_managed" only with claude_request or claude_request_async. The gateway makes a request-scoped --strict-mcp-config from only provisioned gateway-owned local MCP definitions; dynamic npx, ambient PATH, and Codex-config overrides are excluded. Use approvalStrategy:"legacy" for Codex, Gemini, Grok, Mistral, Devin, and Cursor. They reject mcp_managed before launch because their ambient MCP configuration cannot be isolated. approvalPolicy is valid only for Claude managed requests and has no effect with legacy.
  3. No caller-imposed wallclock timeout; poll every 60 s by default: see Polling Strategy below. When the user requests a 90-second progress cadence, do not poll earlier. Do not cancel jobs for taking too long; cancel only on explicit instruction or hard failure. idleTimeoutMs is separate: a no-output safeguard for incremental providers, and a total-runtime bound for terminal-burst ones (see Idle Timeout and Runtime Cap).
  4. Complete mandatory reviews without a cap. Dispatch review work only through the current local stdio gateway MCP surface. Require the terminal JSON verdict APPROVED_UNCONDITIONALLY, CHANGES_REQUIRED, or BLOCKED_EXTERNAL. On CHANGES_REQUIRED, a qualified result, incomplete evidence, or malformed output, fix or rebut findings with evidence and re-review the same required reviewer. Do not set a review round, turn, token, price, budget, or wallclock limit. A failed or unavailable required reviewer must be retained as BLOCKED_EXTERNAL with its exact error, never treated as approved. Repair and retry it until it returns APPROVED_UNCONDITIONALLY, or stop only for explicit user cancellation or a terminal external blocker.

Read the full file on GitHub · 556 lines

Changes

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.

  1. 10d ago First seen · 556 lines · 75 tokens per session scan A 2399047b5d7e

Subscribe to this mod's changes

async-job-orchestration is a skill published in the GitHub repository verivus-oss/llm-cli-gateway (15 stars, last pushed yesterday), licensed MIT. It adds 75 tokens to every session and 7,427 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.