Borrowing it
Nothing to install: this file belongs to maoxx241/vllm-ascend-workspace. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/maoxx241/vllm-ascend-workspace/main/AGENTS.mdgit clone --depth 1 https://github.com/maoxx241/vllm-ascend-workspaceWrote 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/instructions/maoxx241/vllm-ascend-workspace/agents-md)<a href="https://agentmods.dev/instructions/maoxx241/vllm-ascend-workspace/agents-md"><img src="https://agentmods.dev/badge/instructions/maoxx241/vllm-ascend-workspace/agents-md/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/instructions/maoxx241/vllm-ascend-workspace/agents-md"><img src="https://agentmods.dev/badge/instructions/maoxx241/vllm-ascend-workspace/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.01914 | $0.01914 |
| Opus 5 | $0.00957 | $0.00957 |
| Sonnet 5 | $0.00383 | $0.00383 |
| Haiku 4.5 | $0.00191 | $0.00191 |
Grade A, and why
vllm-ascend-workspace AGENTS.md 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 — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository instructions
Local vllm + vllm-ascend development scaffold. vllm/ and vllm-ascend/ are Git submodules.
This repository provides a remote development substrate first, then vLLM-Ascend skills on top.
Remote development model
Use native client tools for local files and local shell work.
Use remote companion tools for remote endpoints. The remote tools mirror native tool semantics and only add endpoint fields:
| Local tool | Remote tool |
|---|---|
| Read | remote.read |
| Edit | remote.edit |
| Write | remote.write |
| Bash | remote.bash |
| Glob | remote.glob |
| Grep | remote.grep |
| LS | remote.ls |
| Monitor | remote.monitor |
| apply_patch | remote.apply_patch |
Default endpoint fields:
hostportuser, defaultrootroot, default/cwd, default/vllm-workspace
Prefer host + port direct endpoints for ordinary remote development.
session_id, session_file, and machine remain advanced compatibility
paths for managed VAWS sessions and legacy single-tenant flows.
Prefer remote companion tools for ordinary remote development. Hooks are
permissive by default, and direct endpoints default to full remote-path
permission (root=/). Pass a narrower root explicitly when a task requires
path isolation.
Skills
Repo-local skills live under .agents/skills/. Each has its own SKILL.md with usage, entry points, and routing rules — read that before invoking.
| Skill | Purpose |
|---|---|
repo-init |
Initialize workspace: gh, GitHub auth, submodules, fork topology |
machine-management |
Add / verify / repair / remove a remote NPU machine |
npu-fleet-monitor |
Deploy, start, inspect, restart, or stop the loopback-only NPU monitoring dashboard from its standalone worktree |
session-management |
Create / inspect / remove / group isolated agent sessions (local worktree + remote container + leases) |
remote-toolbox |
Compatibility backend for managed VAWS target/probe/exec/job/sync/service/artifact/cleanup tools |
remote-code-parity |
Sync local working tree to remote container before execution |
modelscope |
Download / resume / status-check / SHA256-verify ModelScope model weights under explicit local directories |
vllm-ascend-serving |
Start / check / stop a vLLM Ascend service on a remote container |
vllm-ascend-benchmark |
Run vllm bench serve benchmarks (single-run or multi-run with warmup) |
ascend-memory-profiling |
Profile HBM memory usage on Ascend NPU for vLLM serving scenarios |
ascend-profiling-collection |
Collect one Ascend torch-profiler case end-to-end (start service, bracket workload with /start_profile + /stop_profile, run analyse(), verify outputs, write manifest) |
ascend-profiling-analysis |
Analyze collected Ascend torch-profiler roots/manifests and generate reports |
curate-workspace-knowledge |
Explicitly review, deduplicate, promote, merge, reject, or deprecate verified knowledge candidates |
vllm-ascend-graph-debug |
Diagnose Ascend graph compile, capture, replay, and graph/eager correctness divergence |
vllm-ascend-correctness-validation |
Plan and compare baseline/candidate, eager/graph, offline/online, and task-metric correctness evidence |
vllm-ascend-change-validation |
Map code diffs to required validation evidence and aggregate downstream runs into PR reports |
vllm-ascend-performance-regression |
Control alternating baseline/candidate experiments and assess performance regressions |
vllm-ascend-distributed-debug |
Diagnose rank topology, process-group, endpoint, collective, and distributed hang failures |
ascend-operator-debug |
Reduce a model symptom to one Ascend operator call and validate an explicit input/mode matrix |
ascend-triton-operator-development |
Convert a PyTorch or GPU Triton contract into a first correct Ascend Triton candidate |
ascend-triton-kernel-validation |
Detect fallback and validate an Ascend Triton kernel over an explicit correctness matrix |
ascend-triton-kernel-optimization |
Run profiler-driven, correctness-gated Ascend Triton optimization experiments |
ascend-triton-workflow |
Orchestrate Ascend Triton development, validation, optimization, and evidence linking |
vllm-ascend-pd-serving |
Orchestrate grouped prefill/decode services, connector configuration, rollback, and smoke tests |
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 Changed · +14 lines · +302 tokens per session 4c00966af76b
- 9d ago First seen · 99 lines · 1,612 tokens per session scan A a81138f455d4
vllm-ascend-workspace AGENTS.md is an instructions file published in the GitHub repository maoxx241/vllm-ascend-workspace (36 stars, last pushed 4d ago), licensed MIT. It adds 1,914 tokens to every session, about $0.0096 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 instructions, from other repositories
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.