Ascend/agent-skills

40Stars on the repository
91Mods indexed here, across every type
4mo agoLast push, which is what freshness is scored on
MulanPSL-2.0Licence, which decides whether bodies are shown

Ascend/agent-skills

Skill Claude CodeCodex

A guide for managing SSH network tunnels, including local and remote port forwarding and proxy modes. SSH is a secure way to connect to another computer over a network.

not rated 40 +1 4mo ago A 30 tokens MulanPSL-2.0

hccl-test

50

Ascend/agent-skills

Skill Claude CodeCodex

HCCL (Huawei Collective Communication Library) performance testing for Ascend NPU clusters. Use for testing distributed communication bandwidth, verifying HCCL functionality, and benchmarking collective operations like AllReduce, AllGather. Covers MPI installation, multi-node pre-flight checks (SSH/CANN version/NPU…

not rated 40 +1 4mo ago C 69 tokens MulanPSL-2.0

k8s-check-fix

51

Ascend/agent-skills

Skill Claude CodeCodex

A guide for checking the health and security of Kubernetes clusters and repairing confirmed problems. Kubernetes is software for running and coordinating containers across servers.

not rated 40 +1 4mo ago C 29 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Codex

Analyze official Megatron-LM commits, PRs, and branch change sets to identify feature evolution, candidate breaking changes, and migration-relevant events. Use when Codex already has a normalized Megatron change set and needs to explain what changed, which new features matter, and which changes should flow into…

not rated 40 +1 4mo ago A 72 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Codex

Track and normalize change requests against the official Megatron-LM repository by branch, PR, commit, commit range, or time window. Use when Codex needs to collect the exact upstream change set before deeper analysis, especially for branch-aware Megatron and MindSpeed migration work, daily/periodic tracking, or…

not rated 40 +1 4mo ago A 79 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Codex

Map migration-relevant Megatron changes onto the official MindSpeed repository by resolving branch alignment, locating affected subsystems, and identifying concrete adaptation points. Use when Codex has structured Megatron change events and needs to decide whether MindSpeed already covers them, which MindSpeed files…

not rated 40 +1 4mo ago A 72 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Codex

Generate migration deliverables for bringing relevant Megatron changes into MindSpeed after branch alignment and impact mapping are complete. Use when Codex already has a confirmed MindSpeed-to-Megatron branch pairing and needs to produce a migration report, candidate patch, or guarded workspace edits instead of…

not rated 40 +1 4mo ago A 69 tokens MulanPSL-2.0

analyse-coverage

56

Ascend/agent-skills

Skill Claude CodeCodex

A test-coverage analysis aid that looks for areas of code not adequately exercised by automated tests. It produces a coverage analysis report.

not rated 40 +1 4mo ago A 17 tokens MulanPSL-2.0

code-comprehension

57

Ascend/agent-skills

Skill Claude CodeCodex

A code-reading and summarising aid that explains unfamiliar software at several levels, from individual functions to modules and whole systems. It is intended for complex codebases such as machine-learning and distributed-training frameworks.

not rated 40 +1 4mo ago A 61 tokens MulanPSL-2.0

coverage

58

Ascend/agent-skills

Skill Claude CodeCodex

Use when working with coverage.

not rated 40 +1 4mo ago B 7 tokens MulanPSL-2.0

generate-unit-test

59

Ascend/agent-skills

Skill Claude CodeCodex

A unit-test generation aid for Ascend projects. Unit tests are small automated checks that verify individual pieces of code, but the description gives no further detail about what it generates.

not rated 40 +1 4mo ago A 0 tokens MulanPSL-2.0

pytest-writer

60

Ascend/agent-skills

Skill Claude CodeCodex

A helper for writing and improving pytest tests, where pytest is a Python tool for checking that code behaves as expected. It focuses on test files, fixtures, parameterized tests, assertions, debugging, refactoring, and coverage analysis.

not rated 40 +1 4mo ago A 66 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Claude CodeCodex

A test-running helper for MindSpeed-LLM, a large-language-model training project. It can run the project's tests and scan code coverage, which measures how much code the tests exercise.

not rated 40 +1 4mo ago A 36 tokens MulanPSL-2.0

unittest-writer

62

Ascend/agent-skills

Skill Claude CodeCodex

A Python unittest assistant for writing and improving unit tests. Unit tests check small pieces of code in isolation and can include setup, cleanup, assertions, and command-line execution.

not rated 40 +1 4mo ago A 80 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Claude CodeCodex

A configuration migration helper for mapping training settings to the MindSpeed-MM FSDP2 YAML format. YAML is a human-readable configuration file format, and FSDP2 is a distributed model-sharding setup.

not rated 40 +1 4mo ago A 57 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Claude CodeCodex

A migration helper for changing data preparation and loading contracts to MindSpeed-MM FSDP2, a distributed training setup. It covers dataset registration, preprocessing reuse, batch collation, and compatible input fields.

not rated 40 +1 4mo ago A 55 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Claude CodeCodex

An orchestrator for moving new models to MindSpeed-MM FSDP2, a distributed training setup. It coordinates the model, data, configuration, and verification parts of the migration.

not rated 40 +1 4mo ago A 50 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Claude CodeCodex

A model migration helper for moving models to the MindSpeed-MM FSDP2 registration and loading contracts. It addresses model plugins, loading signatures, token and embedding updates, and forward-pass compatibility.

not rated 40 +1 4mo ago A 53 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Claude CodeCodex

A verification helper for MindSpeed-MM FSDP2 migrations. It checks whether model, data, and configuration changes work together in a distributed end-to-end run and preserves evidence of the result.

not rated 40 +1 4mo ago A 59 tokens MulanPSL-2.0

modelscope-cli

68

Ascend/agent-skills

Skill Claude CodeCodex

A command-line tool for downloading models and datasets from ModelScope, a platform for sharing machine-learning models and data. It also checks files, counts model parameters, and diagnoses network problems.

not rated 40 +1 4mo ago D 50 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Codex

Triage a daily msverl regression run by reading the baseline comparison log, stopping on success, extracting the most relevant training failure evidence from the daily training log when needed, collecting recent commits from verl main and MindSpeed master, and ranking the most likely culprit commits with concise…

not rated 40 +1 4mo ago A 68 tokens MulanPSL-2.0

Ascend/agent-skills

Skill Claude CodeCodex

A review helper for adapting GPU code to Huawei Ascend NPU hardware. It examines deep-learning and model-inference code, identifies compatibility problems, and produces a Markdown review report with validation guidance.

not rated 40 +1 4mo ago A 168 tokens MulanPSL-2.0

npu-smi

71

Ascend/agent-skills

Skill Claude CodeCodex

Huawei Ascend NPU npu-smi command reference. Use for device queries (health, temperature, power, memory, processes, ECC), configuration (thresholds, modes, fan), firmware upgrades (MCU, bootloader, VRD), virtualization (vNPU), and certificate management.

not rated 40 +1 4mo ago A 65 tokens MulanPSL-2.0

python-refactoring

72

Ascend/agent-skills

Skill Claude CodeCodex

A Python refactoring guide for finding code smells, improving readability, and applying suitable design patterns. Refactoring means changing code structure without changing what the code is meant to do.

not rated 40 +1 4mo ago A 142 tokens MulanPSL-2.0

At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: