category-name

name	CATEGORY_NAME
description	ONE-LINE SUMMARY OF THE CATEGORY — e.g. "Navigate the symbol graph — callers, callees, definers, impact."
allowed-tools	["mcp__mycelium__CAPABILITY_A","mcp__mycelium__CAPABILITY_B"]
category	<navigation | analysis | quality | operations>
icon	<single emoji>
marketplace_examples	[{"query":"Natural language question a developer would ask","tool":"mcp__mycelium__CAPABILITY_A"},{"query":"Second representative example","tool":"mcp__mycelium__CAPABILITY_B"},{"query":"Third representative example","tool":"mcp__mycelium__CAPABILITY_A"}]

CATEGORY_NAME — short title

When to invoke this Skill

A 2–4 sentence description of the situations that bring the agent to this category. Be concrete:

• "Use when the user asks structural-navigation questions ('what calls X', 'what does Y impact')."
• "Do NOT use for graph-statistics questions — see the graph-analysis Skill for that."

Quick examples

Developer question	Tool
"Natural language question a developer would ask"	mcp__mycelium__CAPABILITY_A
"Second representative example"	mcp__mycelium__CAPABILITY_B
"Third representative example"	mcp__mycelium__CAPABILITY_A

Capabilities under this umbrella

CAPABILITY_A — short imperative description

When: 1–2 sentences on the exact trigger.

MCP invocation:

mcp__mycelium__CAPABILITY_A({
  arg1: "...",
  arg2: ...    // optional; default = ...
})

Equivalent CLI (human reading):

mycelium CAPABILITY_A --arg1 ... --arg2 ...

Result shape:

{ "field1": "...", "field2": [...] }

Typical follow-ups: which tools the agent reaches for next.

CAPABILITY_B — short imperative description

(Same structure as above.)

Parity contract

This Skill covers the listed capabilities per RFC-0090 §Invariants:

• CLI ↔ MCP for each listed capability is byte-identical (name, description, args, JSON output).
• tests/parity.test.json asserts at least one input pair per capability.

Cross-references

• Related Skills: <other-category> (for…)
• Source RFC: e.g. RFC-00XX defining these capabilities