[Skill Submission] every-social-style

[everyskill-bot]

New Skill Submission

Skill: every-social-style
Submitted by: Austin Tedesco
Reason: Every's complete social media style guide for X and LinkedIn. Used daily by the growth team to write on-brand social content. Includes data-backed tactics, post archetypes, voice calibration, and platform-specific patterns.

---

This PR was auto-generated from skills.every.to (agent-api).

AI security review will run automatically.

[cloudflare-workers-and-pages]

Deploying everyskill with    Cloudflare Pages

Latest commit:	bfd910a
Status:	✅  Deploy successful!
Preview URL:	https://78d49527.everyskill.pages.dev
Branch Preview URL:	https://skill-every-social-style-177.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 comprehensive social media style guide for a brand called 'Every'. It contains detailed writing guidelines, voice/tone instructions, content strategy, platform-specific advice for Twitter/X and LinkedIn, and data-backed recommendations. There are no prompt injection attempts, no hidden instructions to override system prompts, no encoded commands, no tool usage (let alone dangerous tool usage), no network calls or data exfiltration vectors, no attempts to read credentials or sensitive files, and no social engineering or manipulation beyond standard brand marketing guidance. The content is entirely focused on legitimate social media copywriting style and strategy."}

GPT-5.2 Analysis

{"verdict":"PASS","issues":[],"reasoning":"Reviewed content is a brand/social media style guide. It contains no prompt-injection patterns (no hidden/encoded directives, no 'ignore previous instructions', no attempts to override system/developer policies), no tool-use instructions (no Bash/terminal commands, no file/system access, no network calls), and no data-exfiltration mechanisms (no credential harvesting, external URL callbacks, or requests to read secrets). It also does not include social engineering aimed at acquiring sensitive information; the guidance is limited to tone/engagement for public social posts. No allowed-tools configuration is present, so nothing is overly permissive."}

---

Frontier model review complete. Human approval still required.