The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the execution of the asc CLI tool and the jq utility to perform operations like listing apps, groups, and testers, as well as importing/exporting data.
[CREDENTIALS_UNSAFE]: The skill explicitly instructs the agent to use a sensitive private key path ~/.asc/AuthKey.p8 for authentication (asc auth login). Accessing private keys is a high-risk operation, although it is a standard requirement for the asc tool's functionality.
[INDIRECT_PROMPT_INJECTION]: The skill implements an 'affordances' pattern where it tells the agent to 'Copy affordance commands directly' from the JSON output of the asc tool. This creates a vulnerability surface where poisoned data returned by the tool (e.g., from a compromised developer account or shared group) could lead the agent to execute unintended or malicious commands.
Ingestion points: Tool output JSON (specifically the affordances field) and testers.csv files used for bulk imports.
Boundary markers: None identified; instructions explicitly tell the agent to trust the affordance commands.
Capability inventory: Subprocess execution via asc, jq, and shell redirections.
Sanitization: None mentioned for tool output or CSV content.
[INDIRECT_PROMPT_INJECTION]: The skill processes testers.csv files for bulk operations (asc testflight testers import). Malicious content within the CSV file could potentially influence the agent's behavior during the import process.

asc-testflight