The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized access behaviors were detected. The skill is consistent with its stated purpose of assisting in the development of Elastic integrations.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, such as user-provided API documentation and sample log files. While this creates a vulnerability surface, the skill implements measures to handle data securely.
Ingestion points: The skill ingests user-supplied API documentation URLs, sample data files (e.g., @samples/acme_event.json), and research briefs as described in SKILL.md and references/add-datastream-workflow.md.
Boundary markers: The skill instructs the agent to wrap specialized subagent instructions in embedded guidance files (e.g., builder-subagent-guidance.md) to establish task boundaries.
Capability inventory: The skill utilizes the elastic-package CLI for filesystem operations, package scaffolding, building, and running system tests in Docker containers.
Sanitization: A mandatory 'Data anonymization' policy is included in references/create-workflow.md and references/add-datastream-workflow.md, requiring all committed data, including test fixtures and sample events, to be fully anonymized.

create-integration