Growth Autopilot

Automate full-funnel strategy generation, budget structure design, and dynamic bid/scale adjustments for Meta (Facebook/Instagram), Google Ads, TikTok Ads, Y...

Security Scan

This is an instruction-only skill that consistently defines policy-generation and decision-rule outputs for ad autopilot behavior and does not request installations, credentials, or external endpoints — the pieces align with a strategy/blueprint generator rather than an automated account connector.

Purpose & Capability

Name, description, and SKILL.md consistently describe a policy/strategy generator for paid growth across ad platforms. The skill does not claim to perform platform API actions and does not require platform credentials, which is proportionate for a policy/blueprint-focused skill.

Instruction Scope

Runtime instructions are limited to generating objectives, policies, decision rules, YAML examples, and pseudocode. They do not instruct reading system files, environment variables, or contacting external endpoints, nor do they grant the agent open-ended permission to gather arbitrary context.

Install Mechanism

No install spec and no code files are provided (instruction-only). Nothing is written to disk or fetched at install time, which is low-risk and consistent with the stated purpose.

Credentials

The skill declares no required environment variables, credentials, or config paths. That is coherent for a policy generation skill; it also means actual integration with ad platforms would require separate connector components not provided by this skill.

Persistence & Privilege

always:false and default model invocation settings are used. The skill does not request persistent presence or system-wide configuration changes, and it does not attempt to modify other skills or agent settings.

Assessment

This skill is a coherent policy/strategy authoring tool — it generates autopilot blueprints and decision rules but does not itself connect to ad platforms or ask for credentials. Before using it in production, ensure you: (1) do not hand the generated policies to an agent or integration that has unrestricted write access to your ad accounts without strict guardrails; (2) provision platform API credentials only to vetted connector components, with least privilege and rate/volume limits; (3) test generated policies in a sandbox or low-budget environment first; (4) enforce logging, auditable change history, and human-in-the-loop approvals for destructive actions (budget freezes, large bid changes); and (5) be aware that absence of code/scan findings only means there is nothing to analyze here — risk arises when you combine this skill with connectors or grant it credentialed access.

Purpose

Core mission:

• Auto-generate full paid growth strategy from goals.
• Auto-design budget and account structure.
• Dynamically adjust bids and scale pace by performance signals.
• Keep growth stable with guardrails and anomaly recovery rules.

When To Trigger

Use this skill when the user asks for:

• automated growth strategy orchestration
• auto budget split and dynamic optimization
• autopilot decision loops for bidding and scaling
• continuous monitoring and adjustment policies

High-signal keywords:

• autopilot, automation, growth ai, growthbot
• budget, bidding, allocation, optimize, scale
• roas, cpa, revenue, performance, campaign

Input Contract

Required:

• north_star_goal
• budget_constraints
• platform_scope
• control_limits (max drawdown, min roas, etc.)

Optional:

• warm_start_data
• creative_inventory_state
• seasonality_rules
• escalation_contacts

Output Contract

1. Autopilot Strategy Blueprint
2. Budget and Structure Policy
3. Dynamic Bid/Scale Rules
4. Safety Guardrails and Kill-switches
5. Monitoring and Escalation Workflow

Workflow

1. Convert business goal to machine-actionable policy set.
2. Initialize budget and structure by channel role.
3. Apply adaptive bid and scale rules by KPI trend.
4. Enforce guardrails and automatic rollback logic.
5. Emit periodic optimization reports and next actions.

Decision Rules

• If KPI drift exceeds tolerance, shift into conservative mode.
• If confidence is low, reduce automation aggressiveness.
• If anomaly severity is high, trigger partial or full freeze.
• If recovery is confirmed, resume staged scale progression.

Platform Notes

Primary scope:

• Meta (Facebook/Instagram), Google Ads, TikTok Ads, YouTube Ads, Amazon Ads, Shopify Ads, DSP/programmatic

Platform behavior guidance:

• Autopilot rules should be channel-specific but policy-governed centrally.
• Keep bid logic aligned with platform optimization objective.

Constraints And Guardrails

• Do not auto-approve risky policy-sensitive creative changes.
• Keep manual override path always available.
• Every auto action must map to an auditable rule.

Failure Handling And Escalation

• If critical metrics are delayed, pause automated changes.
• If policy rejection rate spikes, route to human review queue.
• If data quality degrades, switch to monitoring-only mode.

Code Examples

Autopilot Policy YAML

objective: maximize_revenue_with_roas_floor
roas_floor: 2.3
cpa_ceiling: 38
budget_step_pct: 12
rollback_trigger:
  roas_drop_pct: 18
  window_days: 3

Decision Loop Pseudocode

if roas >= roas_floor and cpa <= cpa_ceiling:
  increase_budget(step_pct)
elif roas < roas_floor:
  decrease_budget(step_pct)
  tighten_bids()

Examples

Example 1: Autopilot bootstrap

Input:

• New account with limited baseline

Output focus:

• starter policy set
• safe exploration bounds
• monitoring cadence

Example 2: Dynamic scale mode

Input:

• KPI stable for 3 weeks

Output focus:

• scale ladder
• bid adaptation rules
• rollback plan

Example 3: Emergency stabilization

Input:

• ROAS crash + spend spike

Output focus:

• freeze/rollback action
• root-cause checklist
• re-entry conditions

Quality Checklist

• Required sections are complete and non-empty
• Trigger keywords include at least 3 registry terms
• Input and output contracts are operationally testable
• Workflow and decision rules are capability-specific
• Platform references are explicit and concrete
• At least 3 practical examples are included