The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill reads API tokens and organization IDs from ~/.axiom.toml. This follows standard security practices for configuring local CLI tools and agents with service credentials.
[COMMAND_EXECUTION]: The skill utilizes a set of Bash scripts to interface with the Axiom API. These scripts leverage curl for networking and jq for robust JSON parsing and construction, preventing common command injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data such as metric names, tags, and values from the Axiom API. The risk of malicious data influencing the agent is mitigated by the skill's use of structured data processing.
Ingestion points: Metrics discovery and query results from Axiom API endpoints.
Boundary markers: Not explicitly implemented, which is typical for discovery-focused skills.
Capability inventory: Network access to Axiom domains and file system access for configuration reading and temporary caching.
Sanitization: Employs jq for all JSON payloads and URL encodes all user-controlled variables in API paths.

query-metrics