The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to execute shell commands using the emdash CLI (and its alias ec) via npx. This includes operations for content CRUD, schema management, and media uploads.
[EXTERNAL_DOWNLOADS]: The skill relies on npx to fetch and run the emdash package from the npm registry. This package is the official management tool provided by the skill author (emdash-cms).
[CREDENTIALS_UNSAFE]: The skill manages sensitive authentication data, such as API tokens and service headers. It utilizes environment variables (EMDASH_TOKEN, EMDASH_HEADERS) and interacts with stored credentials located in a configuration file at ~/.config/emdash/auth.json. This behavior is characteristic of the CLI tool maintaining session state.
[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface related to data processing.
Ingestion points: Content is ingested into the agent context from remote CMS instances through the npx emdash content get and npx emdash content list commands.
Boundary markers: The skill documentation does not provide specific boundary markers or instructions to the agent to disregard potential commands found within the CMS content.
Capability inventory: The agent has access to powerful capabilities, including schema deletion (npx emdash schema delete), content modification, and media management, which could be targeted by instructions embedded in CMS data.
Sanitization: While the CLI converts Portable Text (JSON) to Markdown for easier reading by the agent, it does not explicitly sanitize the resulting text for embedded instructions.

emdash-cli