The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads code from a GitHub repository (Kappaemme-git/codex-first-customer-finder-skill.git) that does not belong to the claimed author ('Aradotso') or a verified organization.
[REMOTE_CODE_EXECUTION]: The installation instructions use npx --yes to download and execute the codex-first-customer-finder-skill package directly from the NPM registry. This allows for the execution of remote code that hasn't been audited during the installation process.
[COMMAND_EXECUTION]: The skill instructions require the user to execute shell commands for manual installation, including git clone, mkdir, and cp, which handle external code and directories.
[PROMPT_INJECTION]: The skill is designed to ingest and analyze data from untrusted external sources, such as arbitrary startup URLs and public GitHub repositories, presenting an indirect prompt injection risk.
Ingestion points: Data is pulled from URLs provided in user prompts (e.g., https://example.com) and external repositories (e.g., https://github.com/username/project).
Boundary markers: The instructions do not specify any delimiters or safety markers to prevent the agent from interpreting content within these external sources as system instructions.
Capability inventory: The skill has the capability to write files to the local file system (./customer-reports) and perform network searches for public signals.
Sanitization: There is no indication of sanitization or validation of the content retrieved from external sources before it is processed by the agent.

codex-first-customer-finder-skill