The Agent Skills Directory

[SAFE]: The skill acts as a standard integration for the Axiom observability platform, allowing users to discover datasets and query metrics via structured bash scripts.
[COMMAND_EXECUTION]: The provided scripts utilize standard system utilities such as curl, jq, and awk to communicate with APIs and process data. User-supplied inputs (like dataset names or metric identifiers) are properly handled using jq for URL encoding and JSON construction, which prevents shell command injection.
[CREDENTIALS_UNSAFE]: The skill identifies the need for API tokens but correctly directs the user to store them in a local configuration file (~/.axiom.toml). The scripts read these credentials locally to include them in the Authorization header for requests to Axiom's official API endpoints, adhering to secure secret management practices.
[EXTERNAL_DOWNLOADS]: Network operations are directed towards Axiom's official domains, specifically api.axiom.co and *.edge.axiom.co. These are verified vendor resources for the Axiom service and are used for legitimate API communication.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API responses (e.g., metric names, tags, and query results). While this constitutes an ingestion surface, the data is processed through scripts that return structured output to the agent environment, minimizing the risk of instructions being accidentally or maliciously interpreted by the LLM.

query-metrics