The Agent Skills Directory

[DYNAMIC_EXECUTION]: The iap-finalizer sub-skill utilizes the Bash tool to execute a local Python utility script (asc.py) using a relative filesystem path (../../_shared/asc-api/asc.py). This script is used to perform GET and POST operations against the App Store Connect API to manage IAP metadata. This execution of local scripts is considered a standard use of functional interfaces for API management.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from WebSearch results (used in originality-check) and local project planning files (used in iap-finalizer, keyword-optimizer, and marketing-strategy). This creates a potential surface for indirect prompt injection where malicious instructions could influence agent behavior or command construction.
Ingestion points: External data from web searches and internal project files like .planning/VALIDATION.md or .planning/MONETIZATION.md.
Boundary markers: The instructions do not specify the use of delimiters or safety prefixes to separate untrusted data from instructions.
Capability inventory: The skill possesses the capability to execute shell commands via Bash and modify project files using Write and Edit tools.
Sanitization: The risk is mitigated by strong process-based controls: the agent is instructed to confirm all data with the user via AskUserQuestion and must perform a mandatory 'dry-run' review, showing the exact command body to the user before using the --apply flag.

app-store