The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read external data such as trace content and metadata using tools like metergraph_get_trace and metergraph_query_traces. This data originates from external application logs and could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Found in SKILL.md via calls to MCP tools that retrieve trace information.
Boundary markers: The skill does not define specific delimiters for untrusted trace content.
Capability inventory: The agent uses MCP tools and standard shell commands to interact with the environment.
Sanitization: No explicit sanitization or filtering of trace content is mentioned before the agent processes it.
[COMMAND_EXECUTION]: The skill describes the execution of local binaries and CLI tools, such as ./bin/start, ./bin/status, and metergraph-cli. These represent standard usage of vendor-provided infrastructure for the MeterGraph platform.
[EXTERNAL_DOWNLOADS]: The instructions involve downloading and verifying "released bundles" and specific package versions like metergraph-cli@0.1.0 from the vendor's official distribution channels.

metergraph-onboarding