Social Growth Ops

Plans, drafts, and schedules social media content (posts, reels, threads) with strict anti-spam and safety rules. Use when the user asks for a content calend...

Security Scan

The skill's requirements and runtime instructions are consistent with a social content planning + scheduling helper; it asks for no credentials or installs and contains clear confirmation guardrails, but provenance is unknown so exercise normal caution before granting posting/automation access.

Purpose & Capability

Name/description match the instructions: task is content calendar, drafts, safety checks, and n8n handoff. The skill requests no unrelated binaries, env vars, or config paths — all required inputs are user-provided content/brand parameters, which is appropriate.

Instruction Scope

Instructions are narrowly scoped to drafting, safety checks, and producing scheduling payloads for workflow runners. Strong guardrails require an explicit user confirmation token (CONFIRM SOCIAL POST) before posting. One minor scope note: it allows the agent to 'ask (or infer from history)' for brand voice inputs — inferring from conversation history gives the agent broader discretion to use prior context; if you want to limit that, instruct the agent to ask instead of inferring.

Install Mechanism

Instruction-only skill with no install spec and no code files. This is low risk because nothing is written to disk and no external packages are pulled.

Credentials

The skill requires no environment variables, credentials, or config paths. That is proportionate to its stated purpose of producing drafts and handoff payloads. It explicitly forbids revealing secrets and forbids auto-actions without confirmation.

Persistence & Privilege

always:false (normal). disable-model-invocation:false (default autonomous invocation allowed) — this is the platform default and not a problem by itself, but remember autonomous invocation increases blast radius if the skill were later modified or combined with credentials.

Assessment

This skill appears coherent and low-risk: it drafts content, runs safety checks, and emits scheduling payloads but does not auto-post without the explicit 'CONFIRM SOCIAL POST' token. Before installing, confirm you are comfortable with the agent using conversation history to infer brand inputs (or instruct it to always ask). Because the skill's source/homepage is unknown, avoid providing account credentials or API tokens to the skill itself — instead keep posting credentials in your workflow runner (n8n) and only hand the agent draft IDs and publishing instructions after manual review. If you plan to wire this to automation, verify the produced n8n payloads offline and test idempotency behavior in a safe environment before granting posting rights.

Purpose

Turn the agent into a reliable “content operator”:

• Content calendar planning
• Post drafting in your brand voice
• Safety checks (anti-spam, duplication, platform constraints)
• Scheduling handoff to n8n/workflow runners

When to Use

Use this skill when the user requests any of:

• A social content plan/calendar
• Captions for TikTok/Instagram/YouTube Shorts/X posts
• Hashtags, titles, hooks, CTAs
• “Repurpose this into 10 posts” / cross-post pack
• A posting schedule and batching strategy
• Automation instructions for n8n workflows

Non-Goals / Safety Limits

• No auto-follow / no auto-DM / no account actions without explicit confirmation.
• No posting without the user confirming the exact post text(s) and target(s).
• No aggressive scraping or bypassing platform protections.
• No financial/crypto actions inside social autopilot.

Brand Voice Inputs

Before drafting, ask (or infer from history) these inputs:

• Audience: (who is it for)
• Offer: (what you sell / promote)
• Tone: (e.g., friendly, expert, edgy, minimalist)
• Language: (RU/EN/mixed)
• Prohibited content: (topics, claims, style)
• Posting frequency: (per day / per week)
• Target platforms: (Telegram channel, X, Instagram, TikTok, YouTube Shorts, etc.)

If inputs are missing, produce a short questions list and stop.

Workflow

1) Build a Content Calendar

Generate a table with:

• Date/time slot (relative if exact timezone unknown)
• Platform
• Content type (post/reel/thread/short)
• Hook style (1 sentence)
• Goal (engagement, lead capture, authority)
• Asset needed (image/video/script/link) or “text only”

Output: calendar.md-style structure.

2) Produce a Draft Pack

For each slot, draft:

• Primary text (caption/poem/script)
• 3 hook variants (for A/B testing)
• Hashtags (platform-appropriate limit)
• CTA (soft / hard CTA options)
• Repurpose notes: how to reuse for other platforms

Output as numbered drafts with stable IDs.

3) Safety + Anti-Spam Checks

Before scheduling:

• Detect duplicates across drafts and propose rewrites
• Ensure no repeated identical CTA
• Ensure claims are not unverified (recommend soft phrasing)
• Ensure platform length constraints are respected
• Ensure no banned keywords list is violated (if provided)

Output: “pass/fail” with exact reasons if fail.

4) Scheduling / Automation Handoff

If the user approves:

• Provide an n8n workflow trigger spec (inputs only)
• Provide a “payload JSON” format to hand to n8n
• Include idempotency keys to avoid double-posting

If the user does not approve, stop and request confirmation.

Confirmation Contract (Required)

Never proceed to schedule or “post” unless the user explicitly says:

CONFIRM SOCIAL POST

And also provides:

• Target platform(s)
• Exact post text(s) (or the IDs of drafts to post)
• When to publish (or “publish now”)

Output Format

Always respond with the following sections:

1. ## Proposed Plan
2. ## Draft Pack
3. ## Safety Check
4. ## Approval Required
5. ## n8n Handoff Spec (if approved)

Example “User Approval”

User: “CONFIRM SOCIAL POST. Publish drafts 1, 2, 3 to X tomorrow 10:00 and Telegram at 12:00.”

Assistant: creates scheduling payloads and idempotency keys.

Guardrails

• Do not reveal secrets (API tokens, cookies, passwords).
• Do not attempt to bypass rate limits.
• Prefer “compose now, schedule later” with explicit approval.