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/skills/ascend-memory-profiling/SKILL.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/skills/maoxx241/vllm-ascend-workspace/ascend-memory-profiling)<a href="https://agentmods.dev/skills/maoxx241/vllm-ascend-workspace/ascend-memory-profiling"><img src="https://agentmods.dev/badge/skills/maoxx241/vllm-ascend-workspace/ascend-memory-profiling/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/maoxx241/vllm-ascend-workspace/ascend-memory-profiling"><img src="https://agentmods.dev/badge/skills/maoxx241/vllm-ascend-workspace/ascend-memory-profiling.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 | $0.00110 | $0.04208 |
| Opus 5 | $0.00055 | $0.02104 |
| Sonnet 5 | $0.00022 | $0.00842 |
| Haiku 4.5 | $0.00011 | $0.00421 |
Grade A, and why
ascend-memory-profiling 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 6d 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 — 290 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Ascend Memory Profiling
Collect and analyze HBM memory usage on Ascend NPU devices running vLLM serving workloads. Produces a structured breakdown of memory by component, with every value traceable to its data source.
Remote substrate rule: use .remote-dev remote tools for ad hoc remote
read/edit/bash/search/patch work around memory profiling setup and output
inspection. Use this skill for the domain HBM workflow and keep its scripts as
the compatibility backend for managed VAWS sessions.
Use this skill when
- the user asks to profile or analyze GPU/NPU memory (显存) usage
- the user wants to understand what consumes HBM in a vLLM serving scenario
- the user asks "权重/KV cache/HCCL/激活各占多少显存"
- the user wants to verify memory allocation against theoretical expectations
- the user asks to compare memory usage across different configurations
Do not use this skill when
- the task is performance profiling (kernel timing, bubble analysis, step/layer/operator breakdown, cross-rank diagnosis) → use
ascend-profiling-analysis(consumes anascend-profiling-collectionmanifest or a remote profile root) - the task is starting/stopping a service without memory analysis → use
vllm-ascend-serving - the task involves non-Ascend hardware
- the task is offline (non-serving) inference only
Data source priority
| Priority | Source | Role | Trustworthiness |
|---|---|---|---|
| P0 | msprof --application wrapping |
Full component breakdown (APP, HCCL, RUNTIME, SLOG) | Highest -- sees memory torch cannot manage |
| P1 | npu-smi info |
Static baseline + phased delta | High -- hardware-level |
| P2 | vLLM startup logs | Weights, KV cache, num_gpu_blocks | Medium-high -- application-reported |
| P3 | safetensors file headers |
Tensor shapes, dtypes, byte sizes (byte-accurate); component classification and shard strategy are rule-based inference | High for byte sizes; medium for per-device attribution |
| P4 | Model config.json |
Theoretical weight calculation (fallback) | Reference only |
What ships with it
7 files 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.
- 6d ago Changed · +8 lines 519d25f9a82d
- 10d ago First seen · 282 lines · 110 tokens per session scan A b0f7896bedaf
ascend-memory-profiling is a skill published in the GitHub repository maoxx241/vllm-ascend-workspace (36 stars, last pushed 5d ago), licensed MIT. It adds 110 tokens to every session and 4,208 once invoked, about $0.0006 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
media-ingest
Ingest video, audio, PDF, book, screenshot, and GitHub repo content into the brain. Multi-format handling with entity extraction and backlink propagation. Covers video-ingest, youtube-ingest, and book-ingest subtypes.
mem0-oss-to-platform
Plan and then execute a migration of a project from the mem0 open-source / self-hosted SDK (the local Memory class) to the mem0 Platform / hosted / managed SDK (the MemoryClient class). Use this whenever a developer wants to move, switch, or migrate their mem0 usage off OSS/self-hosted to the hosted API — e.g.…
Cortex
Operate Cortex, the LifeOS memory system — the typed Knowledge Archive (People, Companies, Ideas, Research with typed related: links) plus recall of prior work sessions, ISAs, and conversations. Search, add, harvest, develop, ingest, distill, graph-navigate, recall. USE WHEN cortex, knowledge, knowledge base, search…
memory
Use when the user asks to remember, recall, forget, update, search, or inspect durable OpenSquilla memory, including profile facts in USER.md and long-term notes in MEMORY.md or memory//.md.
ha-data-stores
Map of Hope Agent's local data stores and safe read-only query workflow. Use when the user asks where Hope Agent stores data, wants to inspect sessions/messages/memory/logs/background jobs/knowledge indexes/settings, asks the model to query local app data, or debugging requires checking persisted state. Trigger…
shellm
Reference for the shellm system — recursive LLM shell, identity management, memory, skills, trajectory, and all CLI tools. Use when working on shellm itself, debugging agent behavior, or understanding how the pieces fit together.