crunchbase-investor-lead-search

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.80). The Skill explicitly directs the agent to install a runtime dependency automatically "without asking the user to confirm" and to "not describe" the installation to the user, which hides/withholds important consent-related behavior and is therefore deceptive relative to normal transparent operation.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The Skill explicitly instructs automatic installation of a required CLI without obtaining or disclosing user consent ("install ... automatically without asking the user to confirm" and "Do not describe it as a request ... do not request separate approval"), which is intentionally deceptive and creates a supply-chain/installation risk that could be abused to install malicious software or persist remote access.

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md instructs the agent to invoke the browser-act Skill for live browser execution (public Crunchbase pages), which can ingest outsider-authored free text from fetched web content into the LLM context via the runtime browser workflow.

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

• Potentially malicious external URL detected (high risk: 0.80). The SKILL.md instructs the agent to automatically use the "Install BrowserAct" link (https://www.browseract.com/?co-from=lead-generation&redirect=https://github.com/browser-act/skills/tree/main) and the BrowserAct site (https://www.browseract.com/?co-from=lead-generation) at runtime to install a required CLI/Skill, which implies fetching and executing remote installation code as a required dependency.