The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to run local Node.js scripts (contract-gen.ts, verify.mjs) and system clipboard utilities (pbcopy, xclip). These operations are used to verify project completion and generate automation strings based on local project data.
[DYNAMIC_EXECUTION]: The agent dynamically constructs shell commands by resolve file paths from user input or filesystem globbing and interpolating them into node execution strings. This allows the skill to target specific project files but introduces a surface for command structure manipulation.
[DYNAMIC_CONTEXT_INJECTION]: The skill instructions include documentation for the !command syntax as a manual fallback for the user. While this syntax can trigger load-time execution in some platforms, its use here is informational and provides a recovery path for permission errors.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from local contract-data.json and contract.md files, which could potentially contain instructions aimed at influencing the agent's behavior during the ideation workflow.
Ingestion points: Local project files in the ./docs/ideation/ path, resolved via glob or direct arguments.
Boundary markers: Absent; the skill does not specify delimiters to separate untrusted file content from system instructions.
Capability inventory: Access to Bash (shell execution), Read (file access), and system clipboard tools.
Sanitization: Absent; the skill relies on internal scripts for data processing without specifying input validation or sanitization rules in the prompt logic.

get-goal-prompt