The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @kaelio/ktx package from NPM and clones the kaelio/ktx repository from GitHub. These resources originate from an external source not identified as a pre-approved trusted vendor.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external data from various sources and feed it into the agent's context, creating a potential vector for indirect instructions.
Ingestion points: Processes content from local markdown files in the docs directory, Notion API pages, and dbt metadata files (manifest.json, catalog.json).
Boundary markers: The skill instructions do not define specific delimiters or instructions for the agent to treat ingested data as untrusted or to ignore embedded instructions.
Capability inventory: The skill provides tools (ktx_sl_search, ktx_wiki_search, ktx_get_metric) that allow agents to search the semantic layer and retrieve business metrics or wiki content derived from these external sources.
Sanitization: There is no evidence of sanitization, escaping, or filtering of the ingested external content before it is presented to the agent.
[COMMAND_EXECUTION]: The installation and initialization process requires the execution of multiple shell commands, including npm install, git clone, pnpm install, and uv sync, which execute code in the local environment.

ktx-ai-data-agents-mcp-context-skills