Skip to content
Back to skills

Progressive Discovery

ASecurity

Progressive disclosure pattern for MCP tools and skills. Teaches agents just-in-time discovery to minimize token usage.

  • 57 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 17, 2026
ai-agentspythongogit

Works with

  • mcp

Security analysis

A100/100

Scanned October 17, 2026

npx -y skills add diegosouzapw/awesome-omni-skill --skill progressive-discovery --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Progressive Discovery?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Progressive Discovery
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/diegosouzapw-progressive-discovery/badge)](https://www.skillsdirectory.com/skills/diegosouzapw-progressive-discovery)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: progressive-discovery
description: "Progressive disclosure pattern for MCP tools and skills. Teaches agents just-in-time discovery to minimize token usage."
category: core
alwaysApply: true
injectionFormat: content
metadata:
  gobby:
    audience: all
    format_overrides:
      autonomous: full
---

## Skills & Discovery

**Skills** — context-aware prompts for common workflows.
- Browse: `list_skills()` on `gobby-skills`
- Load: `get_skill(name="skillname")` on `gobby-skills`
- Search: `search_skills(query="topic")` on `gobby-skills`
- Hubs: `search_hub(query)` on `gobby-skills`

**MCP Tools** — progressive disclosure (each step gates the next):
1. `list_mcp_servers()` — discover servers (~50 tokens)
2. `list_tools(server_name)` — discover tools (~100 tokens per server)
3. `get_tool_schema(server_name, tool_name)` — get full inputSchema (just-in-time)
4. `call_tool(server_name, tool_name, arguments)` — execute

**Core principle:** NEVER load all schemas upfront — 50+ tool schemas consume 30-40K tokens. Fetch schemas only when you're about to call a tool. This reduces token usage by ~95%.

## Common Mistakes

```python
# WRONG — Loading everything upfront (wastes 30-40K tokens)
for server in servers:
    for tool in list_tools(server):
        get_tool_schema(server, tool)  # Unnecessary!

# WRONG — Guessing parameters without schema
call_tool("gobby-tasks", "create_task", {"name": "Fix bug"})  # Wrong param!

# RIGHT — Just-in-time discovery
get_tool_schema("gobby-tasks", "create_task")  # Learn: needs "title" not "name"
call_tool("gobby-tasks", "create_task", {"title": "Fix bug", "session_id": "#123"})
```

## Available Internal Servers

| Server | Purpose |
|--------|---------|
| `gobby-tasks` | Task management |
| `gobby-sessions` | Session handoff |
| `gobby-memory` | Persistent memory |
| `gobby-workflows` | Workflow control |
| `gobby-agents` | Agent spawning |
| `gobby-worktrees` | Git worktrees |
| `gobby-clones` | Repository clones |
| `gobby-merge` | Merge resolution |
| `gobby-hub` | Hub / cross-project |
| `gobby-skills` | Skill management |
| `gobby-metrics` | Usage metrics |
| `gobby-artifacts` | Artifact storage |
| `gobby-pipelines` | Pipeline execution |

Use `list_mcp_servers()` to see all connected servers with live status.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.