The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/pinescript_lint.mjs uses child_process.spawn with the shell: true option to execute a command constructed through simple string concatenation of the PINESCRIPT_LINT_CMD environment variable and command-line arguments. This pattern is vulnerable to shell command injection if file names or arguments contain metacharacters like ;, &, or |.
[DYNAMIC_EXECUTION]: The linter script automatically loads environment variables from various .env files within the project structure (e.g., .opencode/skill/pinescript/.env). This allows a local project file to override the PINESCRIPT_LINT_CMD variable and redirect execution to a malicious binary or command when the linter is invoked.
[EXTERNAL_DOWNLOADS]: The skill defaults to running npx --yes pinescript-lint, which fetches and executes the linter package from the public npm registry at runtime without a pinned version.
[INDIRECT_PROMPT_INJECTION]: The skill processes external Pine Script files and passes their paths to the shell-executing linter script without sanitization. Ingestion points: scripts/pinescript_lint.mjs reads file paths provided via command-line arguments. Capability inventory: Shell command execution via spawn with shell: true. Boundary markers: None implemented. Sanitization: None; arguments are joined directly into the shell command string.

pinescript