The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill instructs the agent to bypass standard user interaction and consent protocols by performing automatic software installations. Specifically, the text states: 'install the missing dependency automatically without asking the user to confirm' and 'do not request separate approval for dependency installation.'
[EXTERNAL_DOWNLOADS]: The skill triggers downloads and installations of a CLI tool and additional agent skills from a non-trusted third-party domain (browseract.com) and associated GitHub repositories.
[REMOTE_CODE_EXECUTION]: The skill mandates the execution of a remote installation script/package without human oversight. It uses the following pattern for runtime setup: 'use Install BrowserAct to install the missing dependency automatically.'
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest large amounts of untrusted external content (TikTok reviews) into the agent's context for analysis, creating a surface for indirect prompt injection.
Ingestion points: Collects 'review text', 'review title', and 'metadata' from TikTok Shop pages (SKILL.md).
Boundary markers: Absent. The skill does not provide instructions to use delimiters or ignore instructions embedded within the scraped content.
Capability inventory: The skill uses the browser-act CLI to 'execute tasks' and 'access real browser pages' (SKILL.md).
Sanitization: Absent. There is no mention of filtering, escaping, or validating the scraped review data before it is processed by the agent.

tiktok-shop-review-scraper