The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/extract.ts utilizes node:child_process.exec to perform media extraction. The implementation interpolates the txid and outputDir variables directly into the shell command string. While there is a length check for the transaction ID, there is no sanitization of shell metacharacters (such as backticks or dollar signs), making the script vulnerable to command injection if an attacker provides a crafted 64-character string or a malicious output directory path.
Evidence: const cmd = outputDir ? txex ${txid} -o "${outputPath}":txex ${txid}`` in scripts/extract.ts.
[EXTERNAL_DOWNLOADS]: The skill's documentation explicitly directs users to download and install a third-party CLI tool (txex) from a public registry using a global installation command.
Evidence: bun add -g txex in SKILL.md.
[REMOTE_CODE_EXECUTION]: The skill facilitates the execution of external, unverified code by shelling out to the txex binary. As this tool is identified as a third-party dependency not maintained within the skill's repository, it presents a supply chain risk where the executed binary could perform unauthorized actions on the host system.

blockchain-media