Welcome to Team n00b

How We Use Claude

Based on usage over the last 30 days:

Work Type Breakdown: Debug Fix ██████████░░░░░░░░░░ 50% Improve Quality ██████████░░░░░░░░░░ 50%

Top Skills & Commands: /goal ████████████████████ 2x/month /effort ██████████░░░░░░░░░░ 1x/month /model ██████████░░░░░░░░░░ 1x/month

Top MCP Servers: computer-use ████████████████████ 77 calls ccd_session ███░░░░░░░░░░░░░░░░░ 11 calls Claude_in_Chrome █░░░░░░░░░░░░░░░░░░░ 1 call

opencode — https://github.com/verbalchainsaw/opencode
- Monorepo with packages/autogoal (the fork actively in use) and the upstream packages/opencode. Treat autogoal as the working tree.

computer-use — desktop UI control for verifying visual fixes. Ask the guide creator for the server config; it's the most-used integration by far.
ccd_session — paired with computer-use for screenshot/observation sessions. Comes online with computer-use.
Claude_in_Chrome — browser automation for live web checks. Low usage but useful for one-off verifications.

/goal — sets the active goal/task context. Use it to scope what Claude should be working on at any moment.
/effort — tunes the effort level (e.g. ultracode for xhigh + workflow orchestration). Reach for it before large audits or migrations.
/model — switches the active Claude model. Use when a task clearly fits a different tier than the current default.

TODO

TODO