The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to save its analytical outputs as markdown files within the {CWD}/deep-insight/ directory on the local filesystem.
[EXTERNAL_DOWNLOADS]: To perform its research functions, the skill instructions include using integrated web search and fetching tools to retrieve data from external URLs provided by the user or found during research.
[PROMPT_INJECTION]: The skill processes untrusted external data (documents, files, and website content). This creates a surface for indirect prompt injection, where an attacker could embed malicious instructions in the analyzed data to influence the agent's behavior.
Ingestion points: Step 1.1 in SKILL.md instructs the agent to read and digest user-provided files and external URLs.
Boundary markers: No explicit boundary markers or 'ignore' instructions for embedded content are defined in the prompts.
Capability inventory: The skill has the capability to write files to the local disk and perform network operations via search tools.
Sanitization: There are no explicit sanitization or validation steps for the content fetched from external sources before it is processed by the agent.

deep-insight