asc-submission-health
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a variety of commands using the
ascCLI tool. These include both read-only diagnostic commands (e.g.,asc validate,asc status) and mutating actions (e.g.,asc submit cancel,asc metadata push,asc web privacy apply). All commands are context-specific to App Store Connect management and include explicit instructions for human confirmation or review.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from local files, specificallyprivacy.jsonand the./metadatadirectory, to update app information on App Store Connect. This introduces a potential attack surface where malicious instructions embedded in these files could influence the agent's behavior during processing.\n - Ingestion points:
privacy.json(referenced inreferences/app-privacy.md) and the./metadatadirectory (referenced inreferences/readiness-repairs.md).\n - Boundary markers: The instructions do not specify technical boundary markers for the data within the files, but they do explicitly require human review of plans and diffs before changes are applied.\n
- Capability inventory: The
ascCLI tool provides extensive capabilities to query, modify, and manage App Store Connect resources, including builds, submissions, and metadata.\n - Sanitization: The skill mitigates risks by instructing the agent to 'Review the planned change before writing' and 'Review the dry-run diff before applying metadata'.
Audit Metadata