html-research-reports

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly instructs the agent to synthesize code, commits, Slack threads and other sources into a self-contained HTML with annotated code snippets and concrete source citations, which creates a high risk the agent will include API keys, tokens, or passwords verbatim if those values appear in the inputs.

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly requires pulling and synthesizing content from web and MCPs (e.g., "synthesize multi-source research (codebase, git history, Slack, web, MCPs)" and "When given access to MCPs (Slack, Linear, git, web), pull from all of them and cite inline"), so the agent will ingest untrusted/user-generated third‑party content and use it to drive report content and decisions.