The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process sensitive external data from the BambooHR API, including time-off requests, benefit details, and employee documents. This creates an attack surface where malicious content embedded in the HR records could potentially influence the agent's behavior.
Ingestion points: Processes data from BambooHR API endpoints (time-off, benefits, files) as described in SKILL.md and references/official-docs.md.
Boundary markers: The skill provides explicit instructions for "Approval Boundaries" and classification of requests to maintain context separation.
Capability inventory: The skill utilizes Read, Glob, Grep, Write, and Edit tools for managing data workflows and local adapters.
Sanitization: Instructions emphasize minimizing data retention and avoiding the exposure of raw employee data in logs or chat transcripts, though specific input sanitization logic for ingested data is not provided.
[EXTERNAL_DOWNLOADS]: The skill references official source code and documentation for BambooHR SDKs hosted on GitHub and the provider's official domain.
Source URLs: Links to official repositories including github.com/BambooHR/bhr-api-python and github.com/BambooHR/bhr-api-php, as well as documentation.bamboohr.com.
Context: The skill includes a specific cautionary note in references/official-docs.md regarding the versioning and availability of the bamboohr-sdk Python package on public registries, advising manual verification before installation.

bamboohr-core-workflow-b