axiom-asc-mcp
Warn
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of a binary from a third-party GitHub repository (
zelentsov-dev/asc-mcp) using the Mint package manager. This repository is not verified as a trusted vendor or well-known service.\n- [INDIRECT_PROMPT_INJECTION]: The skill provides tools for reading App Store reviews (reviews_list), which creates an ingestion surface for attacker-controlled data. This could allow malicious review text to influence the agent's behavior during management tasks.\n - Ingestion points: Untrusted user review content fetched via the
reviews_listtool in the Review Management workflow.\n - Boundary markers: Absent. The skill instructions do not specify any delimiters or safety warnings for the agent when processing review text.\n
- Capability inventory: The skill has significant capabilities, including
app_versions_submit_for_review,beta_groups_add_builds, and managing IAP/subscriptions.\n - Sanitization: Absent. There is no indication that the MCP server or the agent sanitizes or validates the fetched review content before it is incorporated into the context.\n- [COMMAND_EXECUTION]: The setup instructions involve manual shell commands to install dependencies and configure the MCP server with sensitive environment variables, including paths to App Store Connect private keys.
Audit Metadata