The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple shell commands using git for repository synchronization (fetch, pull, stash, remote management), asm for dependency acquisition, and standard filesystem utilities like test, cd, and printf to manage project state and session tracking.
[REMOTE_CODE_EXECUTION]: The skill facilitates the installation of the agent-skill-manager tool via npm and uses it to download and execute other functional skills from the author's GitHub repository (github.com/luongnv89/skills). These sources are recognized as well-known services or vendor-controlled resources and are documented here for transparency.
[DYNAMIC_EXECUTION]: The skill implements dynamic loading by acquiring specialized modules at runtime and invoking them through filesystem paths resolved by the dependency manager based on the requested workflow range.
[INDIRECT_PROMPT_INJECTION]: The orchestrator ingests data from local files like validate.md to control workflow logic, representing an indirect prompt injection surface. Ingestion points: The skill reads idea.md, validate.md, and prd.md from the project directory (SKILL.md). Boundary markers: Present; the skill includes an explicit instruction to treat external file content as data rather than instructions (SKILL.md). Capability inventory: The skill can execute shell commands via asm and git (SKILL.md). Sanitization: Present; the 'Verdict Gate' logic employs strict token matching for specific status words like 'Build it', 'Maybe', or 'Skip it' (SKILL.md).

product-planner