agent-curator
Deep curation of a single agent definition through structured human conversation. Coordinates knowledge research, persona refinement, tooling optimization, and instruction quality improvement. One agent at a time, with due care.
npx skills add majiayu000/claude-skill-registry --skill agent-curator --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Deeply improves an agent definition through a guided, multi-dimension curation process. It coordinates knowledge research, persona refinement, tooling optimization, local corpora materialization, and instruction quality improvements, one agent at a time.
How it works
- Activation: use the snippet "[ACTIVATE:AGENT_CURATOR_V1]" with parameters: an agent path, mode (broad or project), and project name if mode is project.
- Phase 0: Scope Establishment — present a choice between Broad Improvement and Project-Specific Preparation and wait for human selection.
- Phase 1: Initial Analysis — load the agent, assess dimensions with optional numeric scores, notes, and a recommended focus; choose among Continue, Focus on high-gap dimensions, or Skip.
- Phase 2: Knowledge Sources (P0) — audit current sources, identify issues, and choose between researching new sources, addressing issues, or skipping; then proceed to knowledge research and adjudicate each source individually with options to add, modify materialization, or skip.
- Phase 3: Persona & Doctrine (P1) — review current interpretive lens, assess and propose improvements, and decide to accept, keep, or edit.
- Phase 4: Tooling & MCP (P2) — review current tool modes and MCP servers, propose additions or changes, and review recommendations.
- Phase 5: Local Corpora (P3) — evaluate materialization candidates with fields like topic, source, token size, stability, and recommendation; decide to review, accept, or skip.
- Phase 6: Instruction Quality (P4) — report findings with severities (HIGH/MEDIUM/LOW) and propose resolutions.
- Outputs: Updated agent definition and a curation session record.
When to use it
- Activation is intended for ad-hoc agent curation or project preparation; use when an agent definition file exists and a human adjudicator is available for conversation.
What it can touch
- Tools: claude-code is declared as a tool in this skill.
- Outputs include: Updated agent definition and Curation session record.
Caveats
- The process relies on human adjudication during multiple adjudication steps and per-source decisions; execution depends on human input at each step.
- The described workflow includes structured prompts and phase transitions; actual results depend on human choices and the agent's current definition.
# Agent Curator Skill ## Activation Invoke for ad-hoc agent curation or project preparation: ``` [ACTIVATE:AGENT_CURATOR_V1] Parameters: agent: path/to/agent.md mode: broad | project project: {project-name} # Required if mode=project ``` ## Purpose Drive a structured conversation to deeply improve an agent definition across all dimensions: 1. **Knowledge Sources** (P0) — Find and integrate esteemed references 2. **Persona & Doctrine** (P1) — Sharpen identity and mental model 3. **Tooling & MCP** (P2) — Optimize tool and server configuration 4. **Local Corpora** (P3) — Materialize high-value knowledge locally 5. **Instruction Quality** (P4) — Refine priorities, modes, and behaviors ## Two Modes ### Broad Improvement Mode For general agent enhancement: - Changes should improve effectiveness across all use cases - Improvements are candidates for upstream contribution - Long-term maintenance burden considered ### Project-Specific Mode For preparing an agent for a specific project: - Changes may include project-specific knowledge/tooling - Consider whether changes should stay local or go upstream - Immediate utility prioritized, but long-term impact considered ## Execution Flow ### Phas
- Activation
- Purpose
- Two Modes
- Broad Improvement Mode
- Project-Specific Mode
- Execution Flow
- Phase 0: Scope Establishment
- Phase 1: Initial Analysis
- Phase 2: Knowledge Sources (P0)
- Phase 3: Persona & Doctrine (P1)
- Phase 4: Tooling & MCP (P2)
- Phase 5: Local Corpora (P3)
- Phase 6: Instruction Quality (P4)
- Phase 7: Synthesis
What does the agent-curator skill do?
Deep curation of a single agent definition through structured human conversation. Coordinates knowledge research, persona refinement, tooling optimization, and instruction quality improvement. One agent at a time, with due care.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill agent-curator --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.