The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The Python scripts scripts/create_chat.py and scripts/run_workflow.py disable SSL certificate verification by setting verify=False in network calls to Huawei Cloud's IAM and OptVerse services. This makes the skill vulnerable to Man-in-the-Middle (MITM) attacks that could lead to the theft of IAM authentication tokens and sensitive project data.
[CREDENTIALS_UNSAFE]: The scripts/create_chat.py script saves IAM authentication tokens to a plaintext file named optverse_iam_token.txt in the system's temporary directory (/tmp or %TEMP%). On shared or multi-user systems, these tokens could be accessed by unauthorized users to gain control over the user's cloud resources.
[METADATA_POISONING]: The skill's primary documentation in SKILL.md contains a safety claim that tokens are "cached in-memory only (never written to disk)." This is false, as the included scripts/create_chat.py script explicitly implements a disk-based caching mechanism. This discrepancy could lead users or automated systems to underestimate the risk of credential exposure.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted requirement analysis (.md) and model data (.xlsx) files which are used to drive the multi-round decision engine workflow.
Ingestion points: Requirement and data files are uploaded, downloaded, and parsed to construct AI messages.
Boundary markers: The skill lacks explicit delimiters or instructions to ignore embedded commands within these external files.
Capability inventory: The skill possesses significant capabilities, including shell command execution via hcloud and network communication via Python scripts.
Sanitization: Content from these files is not sanitized or validated before being incorporated into the agent's context.

huawei-cloud-optv-solver-assistant