The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill documents the local storage of API credentials in the sensitive file path ~/.paragraph/config.json used by the CLI for authentication persistence.
[EXTERNAL_DOWNLOADS]: References the installation of the @paragraph-com/cli package from the npm registry and the use of a remote MCP server transport at mcp.paragraph.com.
[COMMAND_EXECUTION]: Utilizes multiple Bash commands to interact with the Paragraph service, including piping markdown content to the CLI and executing SQL queries against the publication's analytics schema.
[INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection because it ingests and processes untrusted data from external sources.
Ingestion points: The agent reads external content from posts via post get, subscriber metadata through subscriber list, and search results from the Paragraph platform.
Boundary markers: The instructions do not provide explicit delimiters or instructions for the agent to ignore or isolate potentially malicious commands embedded in the retrieved content.
Capability inventory: The skill possesses capabilities to write files, execute bash commands (paragraph, cat, jq), and perform network operations via the CLI, providing a path for injected instructions to trigger actions.
Sanitization: No evidence of sanitization or validation of external content is present in the skill's instructions before the data is processed.

paragraph-cli