[Skill Submission] kate-top-edit

[everyskill-bot]

New Skill Submission

Skill: kate-top-edit
Submitted by: Kate Lee
Reason: This skill includes editorial flags that I look for when I'm reviewing a piece for final publication.

---

This PR was auto-generated from skills.every.to/submit.

AI security review will run automatically.

[cloudflare-workers-and-pages]

Deploying everyskill with    Cloudflare Pages

Latest commit:	30959a0
Status:	✅  Deploy successful!
Preview URL:	https://6f76e4f7.everyskill.pages.dev
Branch Preview URL:	https://skill-kate-top-edit-17709149.everyskill.pages.dev

View logs

[github-actions]

Frontier Model Security Review

Agent	Verdict
Claude Opus 4.6	PASS
GPT-5.2	PASS

Final Verdict: PASS

---

Claude Opus 4.6 Analysis

{"verdict": "PASS", "issues": [], "reasoning": "This skill is a straightforward editorial checklist for top-editing drafts. It defines a clear set of writing quality checks (vague openers, unsourced quotes, AI-generated text patterns, jargon, hedging, etc.) and specifies a simple checklist output format. There are no prompt injection attempts, no hidden instructions, no tool usage (Bash, file access, or otherwise), no network calls or data exfiltration vectors, no social engineering or manipulation attempts, and no overly permissive allowed-tools declarations. The skill does not reference any tools at all — it is purely a text analysis prompt that instructs the model to scan a draft and return editorial suggestions. The mention of searching the 'Every archive' for relevant pieces (item 6) is an editorial instruction to the LLM to leverage its knowledge, not a dangerous external network call. This is a benign, well-scoped editing skill."}

GPT-5.2 Analysis

{
  "verdict": "PASS",
  "issues": [],
  "reasoning": "No prompt-injection patterns (e.g., 'ignore previous', hidden/encoded directives) are present. The skill contains only editorial instructions for generating a checklist from a provided draft, with constrained output formatting. There is no tool invocation, no shell/Bash usage, no filesystem access, no credential handling, and no instructions to make network calls or exfiltrate data. While it asks to 'Search the Every archive' and suggest URLs, it does not specify external endpoints, scraping, or credentialed access, and remains a content-editing requirement rather than a tool-permission escalation. No social engineering or manipulation attempts detected. No overly permissive tool allowances are included."
}

---

Frontier model review complete. Human approval still required.

[whw]

🚀