silbaram
UserSpec-driven development harness for AI coding agents with human-in-the-loop approval gates. Turns one-sentence ideas into approved specs, dependency-aware task graphs, and verified run logs for Claude Code, Codex, and Gemini CLI.
Categories
Indexed Skills (8)
p2a-dev-execution
Use when implementing one ready Plan2Agent task or a bounded batch of independent ready tasks into real code changes and recording each run, without touching planning artifacts.
p2a-harness
Use when turning a one-sentence product idea into a gated Plan2Agent intake, spec, implementation plan, task graph, and review report.
p2a-intake
Use when extracting requirements, assumptions, and clarification questions from a one-sentence Plan2Agent product idea.
p2a-review
Use when reviewing a Plan2Agent spec, implementation plan, or task graph for missing decisions and execution risk.
p2a-spec
Use when converting Plan2Agent intake output and user answers into product and implementation specs.
p2a-task-author
Use when authoring a Gate C task graph draft from a Plan2Agent context bundle so an agent proposes tasks and a human approves them at the gate.
p2a-task-breakdown
Use when splitting an approved Plan2Agent implementation spec into a dependency-aware task graph.
p2a-next
Use when a Plan2Agent user asks what to do next, wants to resume a project, or needs one state-based next action in an agent session.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.