The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using python3 to run local scripts (scrape_blogs.py, scrape_linkedin_posts.py, search_twitter.py). These commands interpolate arguments such as --urls, --profiles, and --query directly from user-provided configuration values. This pattern is vulnerable to command injection if the input values contain shell metacharacters.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes content scraped from external blogs, LinkedIn profiles, and Twitter/X handles which are under competitor (third-party) control.
Ingestion points: Raw data fetched by blog-scraper, linkedin-profile-post-scraper, and twitter-scraper is fed into the synthesis phase.
Boundary markers: None; the instructions in Phase 4 and 5 lack clear delimiters or instructions to ignore embedded commands within the scraped content.
Capability inventory: The skill has the capability to execute shell commands (python3) and write files to the local filesystem (clients/ directory).
Sanitization: There is no evidence of sanitization, filtering, or validation performed on the scraped content before it is processed by the agent.
[COMMAND_EXECUTION]: The documentation provides a persistence mechanism by instructing the user or agent to configure a cron job for recurring execution of the tracking skill.

competitor-content-tracker