Security checks across malware telemetry and agentic risk
This is a disclosed research-report skill that may use web research and save report files, with no hidden code, credential use, or destructive behavior found.
Install this if you want long, source-backed research reports. Before use, note that default PDF delivery may create files on your Desktop and in an analysis archive, and optional Feishu output may publish content into a third-party document system. Verify high-impact business, investment, policy, or technical claims independently.
ChronoPrism is an instruction-only research framework skill. It contains no executable scripts, no package install steps, no bundled dependencies, no background jobs, and no required environment variables or credentials. Expected behavior: It may guide the agent to use available read-only research tools, such as web search or page fetching, when current public information or citations are needed. It instructs the agent to treat fetched web pages, papers, posts, comments, and retrieved documents as untrusted source material, not as instructions to follow. It may generate research briefs, decision memos, competitor analyses, technical evaluations, or PM opportunity scans in chat. It should create or edit files only when the user explicitly asks for a saved artifact. It does not request account access, tokens, local secrets, shell execution, package installation, persistent scheduling, autonomous loops, or account-changing actions. Primary risk: External sources can be incomplete, outdated, biased, or adversarial. Users should verify high-impact claims and treat outputs as decision support rather than definitive proof. Suggested audit context: Expected risk level: low. Expected behavior type: instruction-only research and synthesis. Any finding related to read-only web access or optional user-requested Markdown output is purpose-aligned.
65/65 vendors flagged this skill as clean.
