feed-ft

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The adapter claims "headlines only" yet embeds explicit, out-of-scope instructions to bypass paywalls (e.g., spoof Googlebot UA, use a 'bypass-paywalls-clean' Chrome extension and archive workarounds), which are deceptive and exceed the skill's stated purpose.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). Runtime path: the skill fetches FT RSS (https://www.ft.com/rss/home / /rss/markets) and ingests channel > item fields like title, link, pubDate, and optionally description—these are publisher-provided text from an external third party (Financial Times) that the operating user did not author.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)