cs-build-plan-app
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted content from sales-call recordings and transcripts to generate app plans. This is a vulnerability surface for indirect prompt injection. \n
- Ingestion points: Data is mined from user-provided recordings, voice notes, and rambled text as described in §1. \n
- Boundary markers: A strict human-in-the-loop 'Gate' is implemented in §6a, requiring the agent to narrate the plan back and obtain explicit user consent before proceeding to the write phase. \n
- Capability inventory: The skill has capabilities to execute local scripts, use the codespring CLI for project management, and perform standard file and git operations. \n
- Sanitization: There is no explicit text sanitization, but the mandatory confirmation gate serves as the primary control against accidental obedience to embedded instructions. \n- [REMOTE_CODE_EXECUTION]: The skill utilizes npx @codespring-app/cli to manage projects. While this involves downloading and executing code from the npm registry, the package belongs to the verified authoring organization and the behavior is essential to the skill's function. \n- [COMMAND_EXECUTION]: The skill executes local bash and node scripts within the codespring directory and uses command-line tools like git, grep, and rg. These operations are scoped to project initialization and metadata management.
Audit Metadata