The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingest and acts upon content from project files that could be influenced by external actors. 1. Ingestion points: The agent reads ROADMAP.final.md, PLAN.final.md, and project principles (PRINCIPLES.md, AGENTS.md) which guide its behavior and the briefing of sub-agents. 2. Capability inventory: The skill possesses significant capabilities including file system writes, git branch management (merge, commit, worktree), and the ability to execute project-specific tools. 3. Boundary markers: There is no instruction to use explicit delimiters or "ignore" instructions when processing these documents. 4. Sanitization: The skill does not perform sanitization or validation on the content of the roadmaps before utilizing them in its orchestration logic.
[COMMAND_EXECUTION]: The skill utilizes several shell commands to manage the project environment. 1. Git Operations: Employs git fetch, git worktree add/remove, git log, git merge, and git commit as part of the orchestration and delivery workflow. 2. Project Tooling: The skill instructs the agent to run project-defined formatters, linters, and test suites, which involves the execution of local code and dependencies.

execute-roadmap