The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses execFileSync to invoke git, gh (GitHub CLI), security (macOS Keychain), and stty. These invocations are scoped to the primary purpose of orchestrating development workflows and managing credentials.\n- [INDIRECT_PROMPT_INJECTION]: The skill includes advanced defensive measures in scripts/workflow-contract.mjs that search for malicious instruction patterns (e.g., bypass safety filters) within untrusted issue and PR data, ensuring the agent treats such data as content rather than authoritative instructions.\n- [DATA_EXPOSURE_AND_EXFILTRATION]: Linear API credentials are managed using industry best practices. The skill encrypts the API key at rest (AES-256-GCM) in the user's config directory with restricted permissions (0600) and leverages the macOS Keychain for secure key storage.\n- [PRIVILEGE_ESCALATION]: Use of the macOS security utility is strictly limited to managing the skill's own credentials within the user's keychain, representing a security hardening measure rather than an escalation attempt.\n- [REMOTE_CODE_EXECUTION]: No remote code execution patterns or unverifiable external dependencies were found. The skill relies exclusively on built-in Node.js modules and local scripts provided within the skill package.

ziw-orchestrate