The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The README.md file contains instructions to download installation scripts from the vendor's domain (updater.linkly.ai) to set up the Linkly AI CLI.
Evidence: curl -sSL https://updater.linkly.ai/cli/install.sh | sh and irm https://updater.linkly.ai/cli/install.ps1 | iex.
[REMOTE_CODE_EXECUTION]: The skill uses a remote CLI tool and MCP gateway to access content. The recommended installation methods for these tools involve piped shell execution from remote vendor-hosted scripts.
[COMMAND_EXECUTION]: The skill instructions direct the AI agent to execute shell commands using the linkly CLI for environment detection, document search, and note management.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data from various document formats, creating a surface for indirect prompt injection attacks.
Ingestion points: Document content retrieved from local and cloud sources via the search, list, and read tools (described in SKILL.md).
Boundary markers: The skill includes a 'Best Practices' section explicitly instructing the agent to treat document content as untrusted and to ignore any instructions or commands embedded within document text.
Capability inventory: The skill utilizes CLI commands and MCP tools to perform file system searches, reading operations, and note writing (note_save). It also performs network requests via the Linkly AI cloud gateway when using the --remote flag.
Sanitization: No explicit code-level sanitization is implemented within the skill; it relies on the agent's adherence to the provided boundary instructions.

linkly-ai