ppt-agent

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill contains multiple deliberate patterns that materially enable opaque execution, prompt-injection and supply-chain risks (prohibiting source inspection, auto-running subagents with injected prompts, and template injection), which could be abused for data exfiltration or backdoor behavior.

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

• Third-party content exposure detected (high risk: 0.85). Step 2A 的 ResearchSynth subagent 在运行时会进行“联网检索/抓取网页内容”，其抓取到的公开网页文本会被整理进 search.txt/search-brief.txt 并作为可读文本喂回 subagent 的 LLM 上下文，属于 OUTSIDER（公共 web 内容）→ 间接提示注入风险路径。

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

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