The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is instructed to execute elastic-package CLI commands (format, lint, check, test) to validate integration quality. These are official developer tools provided by the vendor for the stated purpose of the skill.- [INDIRECT_PROMPT_INJECTION]: As a code review tool, the skill ingests and processes external, potentially untrusted integration code and manifest files. This creates a surface where malicious instructions within the reviewed code could attempt to manipulate the agent's output.
Ingestion points: Processes local package directories, manifest files, and PR diffs in SKILL.md (Step 1).
Boundary markers: None explicitly used to delimit the code under review; the agent is instructed to read the full file context.
Capability inventory: Has the ability to execute shell commands via the elastic-package tool and write findings to a local file in the tmp/ directory.
Sanitization: The skill relies on authoritative domain-specific checklists and rules to govern its analysis, but does not implement character escaping or strict schema validation on the raw source code it reads.

review-integration