pm-collect

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). High risk: the skill explicitly instructs crawling and copying raw project files, configuration files and fetched URL contents verbatim into aggregated output, so any API keys, tokens or passwords present in those sources would be included in the LLM's output and thus risk exfiltration.

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

• Third-party content exposure detected (high risk: 0.85). 该 Skill 会从 $ARGUMENTS 中提取 URL 并在“URL 抓取”阶段对这些 URL 进行运行时网页抓取，将抓取到的正文作为可读文本进入 LLM 上下文进行聚合（属于“公共 web 内容 fetched at runtime — pages/articles scraped sites”）。

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly extracts and fetches all http/https URLs from the runtime $ARGUMENTS (e.g., URLs such as https://competitor.example.com/pricing or https://www.productcompass.pm/...) and injects the fetched page content into the collected materials used by the agent workflow, meaning runtime-fetched external content can directly influence subsequent prompts and behavior.