---

description: "First steps with apothem — profile, rules, plans, and the conformity gate."

Getting started

Apothem provides four core capabilities: a shared governance profile, a rules corpus, a structured planning pipeline, and a real-time conformity gate.

The shared profile

Your apothem profile lives at <harness-config-root>/ — the native configuration directory of whichever AI tool you target. For the Claude Code adapter that root is ~/.claude/; for other harnesses see docs/harnesses/<harness>.md. The profile is the single source of truth that every harness adapter derives its native config from:

<harness-config-root>/      (e.g. ~/.claude/ for Claude Code)
    CLAUDE.md               (project instructions — Claude Code's native format;
                             other harnesses use their own, e.g. AGENTS.md, GEMINI.md)
    settings.json           (hook bindings + permissions)
    rules/                  (behavioral rules, always-on or path-filtered)
    agents/                 (persistent agent definitions, flat <name>.md)
    skills/                 (reusable skill templates)
    commands/               (slash commands)
    hooks/                  (hook runtime + message files)
    conformity/             (mechanical pre-emission gate scripts)
    schemas/                (authorship-header + JSON schemas)

First rule

The example below uses the Claude Code adapter path; substitute your harness install root for other adapters.

cat > ~/.claude/rules/my-first-rule.md << EOF
---
scope: always-on
---

# My first rule

Always include a brief summary at the end of every response.
EOF
apothem verify --harness claude-code

First plan

apothem plan-spec --quick my-task

Creates plans/YYYY-MM-DD--my-task.md in the current project.

Conformity gate

Run from the harness install root (~/.claude/ for Claude Code):

python conformity/conformity-gate.py

Runs all mechanical conformity checks and reports pass/fail per bar.