The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl to perform REST and GraphQL operations against the Weights & Biases API. This is the primary mechanism for run management and data retrieval.
[EXTERNAL_DOWNLOADS]: Communicates with official Weights & Biases infrastructure (api.wandb.ai) and references documentation from official sources (docs.wandb.ai, wandb.github.io). These are well-known services relevant to the skill's purpose.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests data from external sources.
Ingestion points: Run metadata, project metrics, artifact names, and reports are fetched from api.wandb.ai and processed by the agent.
Boundary markers: None identified in the instructional templates; the agent processes the raw JSON responses from the API.
Capability inventory: Uses curl for network requests and interacts with temporary files in /tmp/. No arbitrary code execution or filesystem modification tools beyond its primary scope are noted.
Sanitization: No explicit sanitization or instruction-ignoring delimiters are provided for the data returned from the API.

wandb