The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is vulnerable to command injection in Step 1. It instructs the agent to expand glob patterns using the command bash -O globstar -c 'printf "%s\n" <glob>', where <glob> is directly taken from the user-provided $ARGUMENTS. An attacker could provide a glob string containing shell metacharacters (e.g., ;, &&, or backticks) to execute arbitrary commands on the host system.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted markdown content from documentation files using powerful tools like Bash, perl, and jq.
Ingestion points: Markdown files (.md) specified via arguments or discovered via the Glob and find commands (SKILL.md, Step 1 and 2).
Boundary markers: There are no instructions to use delimiters or ignore embedded instructions when reading file content.
Capability inventory: The skill has access to Bash, Write, Read, Glob, Grep, and specialized documentation search tools (mcp__elastic-docs__search_docs).
Sanitization: No sanitization is performed on the file paths or content before they are used in shell command strings or processed by validation logic.
[EXTERNAL_DOWNLOADS]: The skill uses specialized tools to fetch data from docs-v3-preview.elastic.dev and www.elastic.co. These resources are managed by the skill's authoring organization.

docs-validate-code-samples