The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process and execute shell commands provided in the affordances field of JSON responses from the asc CLI tool. This represents a potential indirect prompt injection surface if the tool's output were to be manipulated by an attacker (e.g., via a compromised App Store Connect account).
Ingestion points: JSON responses from asc CLI commands described in SKILL.md.
Boundary markers: No specific delimiters or safety instructions are provided to distinguish between data and executable commands within the tool output.
Capability inventory: Shell execution of commands for managing certificates, bundle IDs, and provisioning profiles.
Sanitization: The instructions do not mention any validation or sanitization of the affordances content before the agent executes them.
[COMMAND_EXECUTION]: The skill documentation provides examples for using the asc CLI tool to perform sensitive operations such as revoking certificates, deleting provisioning profiles, and managing authentication keys stored in ~/.asc/AuthKey.p8.

asc-code-signing