The Agent Skills Directory

[SAFE]: The skill operates on public job data from Jobindex.dk without requiring user credentials or accessing sensitive local files.
[INDIRECT_PROMPT_INJECTION]: The skill processes job listings from an external website, which constitutes a potential injection surface.
Ingestion points: The htmlFetch and apiFetch functions in helpers.ts retrieve content from external Jobindex.dk endpoints.
Boundary markers: The CLI tool formats results into structured JSON, providing clear delimiters between metadata and external content for the agent.
Capability inventory: The skill utilizes the Bash tool to execute a local CLI that performs network GET requests; it lacks capabilities for filesystem modification or arbitrary command execution.
Sanitization: The implementation includes a visibleHtml function in detail.ts that explicitly strips <head>, <script>, and <style> tags from job listings before they are processed by the agent.
[SAFE]: Environmental controls are enforced via the allowed-tools frontmatter, restricting the agent's interaction to the specific, validated CLI tool path.
[SAFE]: The CLI prevents logic abuse by validating all flags with Zod and rejecting unknown arguments that could lead to unfiltered data extraction or parameter injection.

jobindex-search