Skelly

Skelly Skill

Use this skill when working in repositories that use skelly for codebase context.

Quick Rules

• Prefer Skelly artifacts and navigation commands before opening many source files.
• Keep context fresh before graph-based queries.
• Use skelly enrich <target> <description> to store agent-authored symbol descriptions in .skelly/.context/enrich.jsonl.

Workflow

1. Validate context health:
   • skelly doctor
2. Refresh context when needed:
   • First run: skelly init
   • Incremental: skelly update
3. Query graph/navigation to find exact symbols and file positions:
   • skelly symbol <name|id>
   • skelly callers <name|id>
   • skelly callees <name|id>
   • skelly trace <name|id> --depth 2
   • skelly path <from> <to>
4. Add targeted symbol descriptions as you analyze:
   • skelly enrich <target> "<description>"

Enrich Target Formats

Use one of:

• path/to/file.go
• path/to/file.go:SymbolName
• path/to/file.go:123
• Stable symbol id: path|line|kind|name|hash

If target is ambiguous, refine with file:symbol or file:line.

JSON Mode

When machine-readable output is needed, prefer --json on query commands:

• skelly symbol Login --json
• skelly callers Login --json
• skelly update --json

Artifacts To Prefer

• .skelly/.context/manifest.json
• .skelly/.context/symbols.jsonl
• .skelly/.context/edges.jsonl
• .skelly/.context/nav-index.json
• .skelly/.context/enrich.jsonl

Constraints

• skelly enrich depends on generated state; run skelly generate first if state is missing.
• Call graph resolution is heuristic; verify uncertain edges in source before risky edits.

Converted and distributed by TomeVault — claim your Tome and manage your conversions.