The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to fetch content and site data from an EmDash instance, which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Data is retrieved via npx emdash content list, npx emdash content get, and npx emdash site export as described in SKILL.md.
Boundary markers: The instructions do not specify the use of delimiters or clear separation between CMS-provided data and the agent's system prompt.
Capability inventory: The agent has the ability to execute shell commands (npx emdash), write to the local file system (e.g., site backups and type definitions), and interact with remote APIs over the network.
Sanitization: There is no mention of sanitizing or filtering the retrieved CMS content before it is presented to the agent.
[COMMAND_EXECUTION]: The skill relies on the execution of the emdash CLI tool via npx. This is a vendor-provided tool used for legitimate management purposes but represents the execution of external code within the agent's environment.
[EXTERNAL_DOWNLOADS]: The skill refers to cloudflared for handling authentication tokens through Cloudflare Access, which is a recognized and trusted service.
[CREDENTIALS_UNSAFE]: The documentation notes that the CLI may store authentication credentials and custom headers in a local configuration file at ~/.config/emdash/auth.json. This is standard for CLI tools but identifies a sensitive local path where credentials may be exposed.

emdash-cli