The Agent Skills Directory

[SAFE]: The skill is entirely instructional. It provides a structured methodology for auditing and A/B testing mobile application paywalls using industry-standard frameworks (RevenueCat, Superwall, Adapty).
[NO_CODE]: No scripts, executables, or package dependencies are included or referenced in the skill body.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data including project-specific marketing files (app-marketing-context.md) and user-provided subscription metrics.
Ingestion points: Reads app-marketing-context.md and accepts App IDs and metrics as user input.
Boundary markers: None explicitly defined in the instructions.
Capability inventory: No file system writes, network requests, or shell command executions are present in the skill.
Sanitization: None specified.
Risk: Since the skill lacks tool-based capabilities, the potential for indirect prompt injection to cause harm is negligible.
[DATA_EXPOSURE]: The skill requests App IDs and subscription metrics to perform its task. This is consistent with its stated purpose of conversion optimization and does not involve exfiltration to third-party domains or hardcoded credentials.

paywall-optimization