veomni-profile

Use this skill for performance profiling and optimization. Two modes: (1) Analyze existing profile files (Chrome traces, memory snapshots) — write scripts to parse and summarize metrics per user requirements. (2) Generate profiles during development — configure ProfileConfig, run training, collect traces, analyze bottlenecks, and suggest optimizations. Trigger: 'profile', 'performance', 'slow', 'MFU', 'throughput', 'bottleneck', 'memory usage', 'trace', 'optimize training speed'.

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

75%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, actionable skill body with concrete code, config, and command examples plus a clear two-mode workflow and validation step. Splitting the NPU and detailed bottleneck material into reference files would improve progressive disclosure and token economy.

Suggestions

Move the NPU (Ascend) profiling section and the Mode 2 bottleneck/solutions table into a separate references file (e.g., references/npu_and_bottlenecks.md), keeping SKILL.md a lean overview with a clearly signaled pointer.

Show the concrete merge_chrome_trace.py invocation inline in Mode 1 step 4 rather than only naming the script, so the multi-rank guidance is copy-paste ready.

Tighten restated guidance such as 'Write analysis scripts as in Mode 1' by linking back to the earlier example instead of repeating the instruction.

Dimension	Reasoning	Score
Conciseness	The body is dense and efficient with a useful component table and no basic-concept padding; a few steps restate earlier guidance (e.g., 'Write analysis scripts as in Mode 1') that could be trimmed, keeping it just below fully lean.	4 / 5
Actionability	Provides executable Python, a concrete YAML config, bash launch commands, and real file paths; minor gaps such as the merge_chrome_trace.py invocation not being shown inline keep it from being fully copy-paste ready everywhere.	4 / 5
Workflow Clarity	Both modes are clearly sequenced with numbered steps, and Mode 2 includes explicit validation in Step 5 (re-profile, verify correctness, document); Mode 1 is read-only analysis without validation checkpoints, a minor gap.	4 / 5
Progressive Disclosure	Well-organized with clear section headers and an overview table, but no bundle files exist so all content is inline in a ~150-line single file; sections like the NPU guide and bottleneck table could be split into references for cleaner navigation.	4 / 5
	Total	16 / 20 Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, well-structured description that explicitly covers both what the skill does (two profiling modes) and when to trigger it, with concrete actions and natural trigger terms. Minor tightening of a few high-level action phrases would push specificity and trigger quality to full marks.

Dimension	Reasoning	Score
Specificity	Lists several concrete actions across two modes (parse/summarize metrics, configure ProfileConfig, run training, collect traces, analyze bottlenecks, suggest optimizations), with minor high-level phrasing keeping it just below fully comprehensive.	4 / 5
Completeness	Clearly states what the skill does (two modes with concrete actions) and when to use it (explicit Trigger clause with concrete phrases), satisfying both requirements explicitly.	5 / 5
Trigger Term Quality	An explicit Trigger clause provides natural user terms ('profile', 'performance', 'slow', 'MFU', 'throughput', 'bottleneck', 'memory usage', 'trace') with good synonym coverage, missing only a few common variations.	4 / 5
Distinctiveness Conflict Risk	Targets a clear niche (VeOmni training profiling/optimization) with distinct, specialized triggers and minimal overlap with other skills.	5 / 5
	Total	18 / 20 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
referenced_paths_exist	Referenced path issues: 3 missing, 3 deeper-than-1-level	Warning

	Total	15 / 16 Passed

Repository: ByteDance-Seed/VeOmni
Path: .agents/skills/veomni-profile/SKILL.md
Commit: c465a63

Reviewed: about 19 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.