The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it is designed to ingest and process untrusted external data.
Ingestion points: The agent processes manifest.yml, changelog.yml, and Handlebars template files (*.yml.hbs) as part of its core functionality defined in SKILL.md and references/manifest-rules.md.
Boundary markers: The skill does not provide instructions to use XML tags, delimiters, or explicit 'ignore instructions' warnings when reading these external files.
Capability inventory: The skill utilizes the elastic-package CLI tool (lint, check, and changelog management) to process this data, as mentioned in SKILL.md and references/changelog-patterns.md.
Sanitization: No sanitization or escaping of the processed file content is described before the data is analyzed by the agent.
[COMMAND_EXECUTION]: The skill provides detailed instructions for executing the elastic-package CLI utility. This is an official development tool provided by the vendor (Elastic) and is used for linting and validating integration package metadata.

package-spec