Prerequisites

Prerequisite check: Huawei Cloud CLI (hcloud) >= 7.2.2 required Run hcloud version to verify the version is >= 7.2.2, and hcloud configure list to confirm a profile exists. If it is not installed or the version is too low, see references/koocli-installation-guide.md for the installation guide.

hcloud version
hcloud configure list

Prerequisite check: Python >= 3.6 required Run python --version to verify the version is >= 3.6.0 If it is not installed or the version is too low, The skill execution is interrupted proactively, and the user is prompted to install Python 3.6 or later. If the Python version is 3.6.x, run the following command to upgrade pip to the latest version pip3 install --user --upgrade 'pip<22'

⚠️ Mandatory Execution Rules

Rule 1: Step-by-step Confirmation

After each step, you MUST:

Print the completion status of the current step
Show the output produced by that step
Every Skill run starts from Step 1 — do NOT skip Step 1 and jump into later steps
Strictly follow this Skill's rules and wait for user confirmation after every step
Strictly follow the described flow; do not perform extra operations. If anything errors, return the error to the user as-is — do not try to fix it yourself
All execution rules take precedence over efficiency

Do NOT chain multiple steps in a single run!

Rule 2: Fixed Directories

After each step, you MUST:

Save all intermediate files under the data/ directory

Rule 3: Permission Issues

When you hit a permission problem, try sudo. If that does not resolve it, error out and abort the Skill flow

Rule 4: Anti-skip Check

Before executing each step, verify that its prerequisites are satisfied:

Step	Prerequisite
Step 2: Environment check	Step 1 is complete; Huawei Cloud AK/SK have been obtained
Step 3: Container environment collection	Step 2 is complete with no errors; `data/` and `artifacts/` are emptied
Step 4: Metric scoring	Step 3 is complete; `data/cloud-native-collection.md` has been generated
Step 5: Report generation	Step 4 is complete; `artifacts/cloud-native-summary.xlsx` has been generated

Rule 5: Information Collection Method

For every piece of information collected, prefer the Huawei Cloud KooCLI tool (hcloud commands)

Rule 6: Status Tracking

📋 Cloud-Native Container Assessment — Status Board
├────────────────────────────────────────────┤
│ Step 1: Configuration       [⏳ In progress]│
│ Step 2: Environment check   [○ Pending]     │
│ Step 3: Information collect [○ Pending]     │
│ Step 4: Metric scoring      [○ Pending]     │
│ Step 5: Report generation   [○ Pending]     │
└────────────────────────────────────────────┘

After each step, update the status board:

📋 Cloud-Native Container Assessment — Status Board
├────────────────────────────────────────────┤
│ Step 1: Configuration       [✅ Done]       │
│ Step 2: Environment check   [✅ Done]       │
│ Step 3: Information collect [✅ Done]       │
│ Step 4: Metric scoring      [✅ Done]       │
│ Step 5: Report generation   [⏳ In progress]│
└────────────────────────────────────────────┘

🔐 Permission Boundary

The AI is allowed to:

Follow the SKILL steps exactly
Invoke only the scripts specified in the SKILL
Return errors as-is when problems occur

The AI is forbidden to:

Temporarily modify or bypass the workflow
Switch to an alternative approach (e.g. fall back to Markdown when html fails)
Modify any script or configuration without user consent
Perform any operation not explicitly listed in the SKILL

Handling out-of-bounds operations

Any out-of-bounds operation must stop immediately and wait for user confirmation. Continuing is forbidden.

Core Workflows

Step 1: Configuration

Prompt the user for the environment configuration.

Item	Description	Example
Huawei Cloud AK/SK	Access Key and Secret Key for Huawei Cloud API	AK/SK = HPUAN3EWCG... / 1Bt5sdDU..
Region	Region where the container cluster lives	region = cn-north-4
Cluster name	CCE cluster name (Step 1 only)	cce_name = dify-cce-cluster
Dockerfile source	Source code repository (containing the Dockerfile)	Dockerfile = https://github.com/langgenius/dify

Once the user supplies the configuration, save the values into environment variables HWC_AK, HWC_SK, CCE_Region, CCE_NAME, Dockerfile_REPO_URL respectively.

Diamond Gate 1: Get user confirmation before Step 2.

Step 2: Environment Check

Using the configuration provided in Step 1, verify that the environment is reachable and that local runtime dependencies are present.

Use hcloud with AK/SK to access the Huawei Cloud CCE environment and confirm it is reachable, If it is not installed or the version is too low, see references/koocli-installation-guide.md for the installation guide.
Verify that the local Python environment is working correctly, Python version requirement is greater than 3.6, If the Python version is 3.6.x, run the following command to upgrade pip to the latest version pip3 install --user --upgrade 'pip<22'
Check whether the Python dependency library is installed. Go to the references directory. If the Python version is 3.6.x, run the command python3 -m pip install -r requirements.txt. If the Python version is later than 3.7, Execute commands using a Python virtual environment pip3 install -r requirements.txt.
Empty any historical files inside data/; if data/ does not exist, create it
Empty any historical files inside artifacts/; if artifacts/ does not exist, create it
Once the checks complete, return the result to the user and wait for confirmation on whether to install dependencies or to continue

Step 3: Container Environment Information Collection

For every metric listed in references/cloud-native-checklist.xlsx, collect the corresponding environment information.

Every collection run MUST be driven by the actual metric items in references/cloud-native-checklist.xlsx and MUST invoke scripts/collect_all.py to collect fresh data — do NOT reuse historical or cached data
Acceptance metric: the metric item being assessed
Quantified target: the reference standard for that metric
Acceptance method: how the environment information for that metric is collected
During collection, if any required piece of information cannot be obtained, explore alternative ways to retrieve it and ask the user for confirmation
Once collection is complete for every metric, fill the collection method (including, but not limited to, Python scripts and executed commands) and the collected data into templates/cloud-native-assessment-template.md, write the result to data/cloud-native-collection.md, and show it to the user

Step 4: Metric Scoring

Based on the collection data in data/cloud-native-collection.md, score each metric:

Compare the collected data against the quantified target to derive an acceptance verdict (Fully Satisfied / Mostly Satisfied / Partially Satisfied / Not Satisfied; metrics that cannot be compared are recorded as Not Satisfied; "Not Applicable" and "Not Evaluated" both roll up into Not Satisfied, and the basis field must state that the metric is N/A or Not Evaluated) and a scoring basis (describe based on the collected data — include the metric's environment information, the collection method used, and the reason for non-satisfaction)
Convert the verdict to a score: Fully Satisfied = 3, Mostly Satisfied = 2, Partially Satisfied = 1, Not Satisfied = 0
Invoke scripts/score_and_excel.py to produce a fresh scoring sheet, output as Excel containing the columns: number, cloud-native dimension, level, acceptance metric, quantified target, acceptance method, description, acceptance verdict, score, full score, scoring basis. Save to artifacts/cloud-native-summary.xlsx

Step 5: Report Generation

Using the contents of the scoring sheet artifacts/cloud-native-summary.xlsx, generate the final assessment report:

The final report uses templates/report_template.md as its base template
The report title is fixed as "Cloud-Native Assessment Report"
The report contains 4 chapters; Chapter 1 and Chapter 2 reuse the content of templates/report_template.md directly
Chapter 3 of the final report is populated from the scoring sheet artifacts/cloud-native-summary.xlsx
Invoke scripts/make_charts.py against the cloud-native-summary.xlsx produced in Step 4 to generate a radar chart and a staircase chart

Chart generation requirements:
- Radar chart: group metrics by cloud-native dimension. The score for each dimension = (sum of that dimension's metric scores / sum of that dimension's full scores) × 5. The radar has exactly six dimensions: Service-orientation, Security, Automation, Elasticity, Observability, Resilience. If a metric belongs to multiple dimensions (e.g. metric #1 belongs to both Service-orientation and Automation), it contributes to the scoring of every dimension it belongs to.
- Staircase chart: overall score = (sum of all metric scores / sum of all full scores) × 5. Stage thresholds — score ≤ 1: Traditional; 1 < score ≤ 2: Basic Cloud; 2 < score ≤ 3: Service-Oriented; 3 < score ≤ 4: Automated; score > 4: Intelligent. The chart must visually render as stairs.
Place the generated radar chart into report section 3.3.3 and the staircase chart into 3.3.4
Replace Chapter 4 of the final report with the actual remediation recommendations
Based on the assessment results, output remediation recommendations by P0/P1/P2 priority:
- P0: Not Satisfied (0 pts) — mandatory items
- P1: Partially Satisfied (1 pt) — items to fix
- P2: Mostly Satisfied (2 pts) — recommended improvements
Invoke scripts/make_report_html.py to generate the final report as a html, written to artifacts/cloud-native-report.html

Core Commands

Get cluster ID and basic info

hcloud cce ListClusters --cli-region=cn-north-4

Get cluster kubeconfig certificate

hcloud cce CreateKubernetesClusterCert \
  --cli-region=cn-north-4 \
  --cluster_id=a5659ec8-55b5........ \
  --duration=1

Get detailed cluster info

hcloud cce ListClusters --cli-region=cn-north-4

Authentication parameters

Auth mode	Required	Optional
AKSK	`--cli-access-key`, `--cli-secret-key`	`--cli-security-token`
Profile	`--cli-profile`	`--cli-mode`, `--cli-region`

Output Format

The final report is delivered as a html file
The scoring sheet is delivered as an Excel file
All output files are saved under the artifacts/ directory

Verification

Post-install verification

Version check: hcloud version MUST return 7.2.2 or higher
Help check: hcloud --help MUST list the available services

Authentication verification

Profile check: hcloud configure list MUST show the configured profile

Best Practices

Evaluate and analyze the container environment running on Huawei Cloud.
Harden the configuration of existing container clusters and modify the cluster configuration based on the optimization suggestions.

References

report_template.md — Final report template
cloud-native-checklist.xlsx — Description of assessment metrics
requirements.txt — Python dependency library

Scripts

collect_all.py — Main assessment script; integrates all collection modules
make_charts.py — Chart generation script
make_report.py — Report generation script
score_and_excel.py — Scoring sheet generation script

Templates

report_template.md — Final report template
cloud-native-assessment-template.md — Data collection template

Notes

Make sure the Huawei Cloud container environment is reachable
Make sure the Huawei Cloud credentials are valid and not expired
If the Huawei Cloud Koocli tool and Python environment cannot be automatically installed using the skill, it is recommended that users install them manually.
Currently only public-network access to the container environment is supported
Output may be truncated during execution; clearing historical data and re-running the Skill is recommended in that case

huawei-cloud-cce-env-assessment

Huawei Cloud — CCE ENV Assessment Skill

Overview