The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the @useclawlink/cli tool to interact with Affinity.
[EXTERNAL_DOWNLOADS]: The skill utilizes npx to download and run the @useclawlink/cli package from the NPM registry at runtime.
[CREDENTIALS_UNSAFE]: The skill references and accesses a sensitive credential file stored locally at ~/.clawlink/credentials.json for authentication.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection by ingesting untrusted data from Affinity records (e.g., person names, company descriptions, list entries). * Ingestion points: External data is fetched via actions such as affinity_get_all_persons, affinity_get_all_companies, and affinity_get_all_list_entries_on_a_list in SKILL.md. * Boundary markers: The instructions advise the agent to read first and confirm with the user before performing write actions, providing a manual checkpoint. * Capability inventory: The skill can execute CLI commands, perform network operations via the CLI, and write to Affinity via run affinity commands. * Sanitization: No explicit sanitization or filtering of the ingested CRM data is described within the skill instructions.

affinity