The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data from OpenTelemetry registry files (YAML) and templates (Jinja2/MiniJinja) to generate code.
Ingestion points: Files located in telemetry/registry/ and telemetry/templates/ are read and processed.
Boundary markers: There are no explicit instructions to the agent to ignore or delimit potentially malicious instructions within these files.
Capability inventory: The skill utilizes powerful tools including the weaver CLI, docker, git, and various code formatters (gofmt, prettier, ruff).
Sanitization: No specific sanitization or validation of the input file content is described beyond the tool's built-in schema checks.
[COMMAND_EXECUTION]: The skill requires the execution of shell commands for its core functionality.
Evidence: Instructions include weaver registry check, weaver registry generate, docker run, gofmt -w, and git diff --exit-code.
Context: These commands are used for schema validation, code generation from templates, and verifying code freshness in CI/CD environments.
[EXTERNAL_DOWNLOADS]: The skill directs the user or agent to download the Weaver tool from external sources.
Evidence: References to https://github.com/open-telemetry/weaver/releases and the otel/weaver Docker image on Docker Hub.
Context: These are the official distribution channels for the OpenTelemetry project's Weaver tool.

otel-weaver