The Agent Skills Directory

[COMMAND_EXECUTION]: The workflow instructs the agent to execute shell commands that incorporate user-supplied input, specifically LinkedIn URLs (e.g., python3 revops/content/fetch_interactions.py "<post_url>"). If the agent does not properly escape the input, a malicious user could provide a URL containing shell metacharacters to execute arbitrary commands on the host system.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to scrape untrusted user-generated content (comments, reactions, and reposts) from LinkedIn, creating a surface for indirect prompt injection.
Ingestion points: Data is fetched from external LinkedIn posts via fetch_interactions.py and stored in content.db (file: SKILL.md).
Boundary markers: There are no instructions or delimiters defined to isolate the scraped content or warn the agent against following instructions embedded in the data.
Capability inventory: The skill has the ability to execute local Python scripts, perform database operations, and make network requests to third-party APIs (Fullenrich, Attio, Extruct) (file: SKILL.md).
Sanitization: The instructions do not mention any sanitization, filtering, or validation of the scraped text before it is stored or used for further enrichment and CRM synchronization.

post-engagers