high-agency

Plugin high-agency from marketplace high-agency in Ssojux2/high-agency-claude-code. Commands read from the repo's manifest — today.

Skills (4)

• bounded-autonomy

  Use when the user asks Claude to keep working autonomously, iterate until done, fix until tests pass, or run a bounded Ralph-style loop with objective or practically verifiable acceptance criteria.

• high-agency-coding

  Use when implementing, debugging, refactoring, reviewing, or planning code where autonomous execution and evidence-based verification are useful; defaults to single-agent execution and escalates models, effort, verification, or review only when they materially improve correctness or efficiency.

• high-agency-doctor

  Diagnose High Agency installation, hook activation, native subagent routing, model/effort overrides, and missing runtime evidence in Claude Code without changing project files or settings.

• high-agency-unit

  Use when writing or changing one unit of code - a feature slice, bug fix, refactor step, perf fix, or throwaway prototype - to build exactly what the unit needs and nothing more; supplies unit playbooks, an over-implementation checklist applied to the diff, and the unit report returned to high-agenc

Slash commands (1)

• /high-agency:doctor

  Diagnose High Agency routing, hooks, model/effort overrides, and environment without modifying project files

Subagents (6)

• high-agency-advisor

  Use for ambitious codebase-wide strategy, long-horizon planning, cross-system coordination, or difficult review where Fable-level reasoning can guide cheaper execution models. Do not edit.

• high-agency-builder

  Use for an isolated implementation work package when delegation saves main-context cost or enables non-overlapping parallel work. Avoid overlapping writes with other agents.

• high-agency-deep-critic

  Rare Fable escalation for security-critical review, high-impact architecture, long-horizon failure analysis, or a hard unresolved reasoning problem after strong evidence-based attempts have failed. Do not edit.

• high-agency-planner

  Use only for high-leverage architecture, ambiguous cross-system decisions, or complex root-cause reasoning where a stronger independent plan materially reduces implementation risk. Do not edit.

• high-agency-scout

  Use for cheap read-only mapping of an unfamiliar codebase, dependency path, or file surface when broad exploration would bloat the main context. Do not edit.

• high-agency-verifier

  Use for cheap deterministic verification: run targeted tests/checks, inspect exact outputs, and report evidence without editing code.

Hooks (8)

• UserPromptSubmit

  python

• PreToolUse

  python

• PostToolUse

  python

• PostToolUse

  python

• PostToolUseFailure

  python

• PostToolUseFailure

  python

• Stop

  python

• SubagentStart

  python

Similar plugins