Discovering TraceDecay

Treat skills as workflow guidance and the CLI or MCP as interchangeable tool transports. MCP is optional: capability discovery must work from a shell alone.

Discovery ladder

1. Run tracedecay tool to list every tool by category with a one-line summary. Scan the relevant category; do not guess tool names.
2. Run tracedecay tool <name> --help for the selected tool's exact arguments, types, enum values, and examples.
3. Run tracedecay --help for lifecycle commands such as init, sync, doctor, daemon, sessions, analytics, automation, and upgrade. Then inspect one command with tracedecay <command> --help.
4. Choose the smallest matching workflow and continue with its skill:

Intent	Workflow skill
Locate or understand code	tracedecay:exploring-code
Trace calls or assess blast radius	tracedecay:tracing-functions / tracedecay:assessing-impact
Review or edit code	tracedecay:reviewing-changes / tracedecay:editing-safely
Recover decisions or session history	tracedecay:project-memory / tracedecay:managing-session-context
Diagnose adoption or host health	tracedecay:diagnosing-analytics / tracedecay doctor
Fix compiler or type errors	tracedecay:fixing-build-and-type-errors

Rules

• Prefer live --help output over remembered flags; the installed binary is the source of truth.
• Do not query TraceDecay databases directly to discover capabilities.
• Do not dump the entire catalog into the answer. Name the selected workflow, exact command or tool, and why it matches.
• Once a tool is selected, use MCP when available or the identical tracedecay tool <name> CLI surface when it is not. See tracedecay:using-the-cli for argument construction.

Deliverable

Return the chosen skill or workflow, the exact next command/tool, and one-line reason. If no capability fits, say what is missing instead of inventing one.