insforge-cli
Pass
Audited by Gen Agent Trust Hub on May 17, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill provides instructions to download and execute the
flyctlinstallation script from Fly.io's official domain (https://fly.io/install.sh | sh). This is a documented step for users requiring source-mode container deployments. - [COMMAND_EXECUTION]: Extensive use of
npx @insforge/clifor project management, database queries, and service deployments. The skill also shells out toflyctlfor remote container builds when in source mode. - [DATA_EXFILTRATION]: The skill manages sensitive information including
OPENROUTER_API_KEY, Stripe secret keys, andINSFORGE_ACCESS_TOKEN. Authentication credentials are stored locally in~/.insforge/credentials.jsonwith restricted permissions. These operations are standard for a development CLI tool. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it ingests untrusted data from the user (such as SQL queries in
db queryor environment variables incompute deploy) and passes them to powerful capabilities (database execution, shell commands). There are no explicit boundary markers or sanitization steps mentioned in the instructions to mitigate potentially malicious content embedded in these inputs.
Audit Metadata