The Agent Skills Directory

[DATA_EXPOSURE]: The script scripts/evaluate_examples.py contains a hardcoded absolute file path to a local user directory (/Users/jaitoor/Documents/Last30Days). This practice can expose the local directory structure and username of the developer's environment and may cause execution failures in environments where the path does not exist.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and synthesize large amounts of untrusted content from external public sources such as Reddit, X (Twitter), YouTube, and general web searches. This creates an attack surface where malicious instructions embedded in social media posts or search results could potentially influence the agent's behavior.
Ingestion points: External data is fetched from various social media platforms and web endpoints as defined in SKILL.md and references/source-map.md.
Boundary markers: The skill instructions do not explicitly define boundary markers or 'ignore' instructions for the external data being processed.
Capability inventory: The skill has capabilities for network operations (via curl and tools), shell command execution (deepline CLI), and local file writing (via evaluation scripts).
Sanitization: No explicit sanitization or escaping logic is present in the provided scripts for handling external input before it reaches the model context.
[COMMAND_EXECUTION]: The SKILL.md file provides instructions to execute several shell commands, including npm install -g deepline, git ls-remote, and various deepline CLI operations. While these are part of the skill's legitimate functionality, they involve interaction with the host system's shell.

deepline-research