Appendix D: Slash Commands Catalog

Built-in commands

Type / to see the current list in your version. The ones that matter most:

Command	Use
/init	Generate a starter CLAUDE.md
/memory	Edit memory files
/clear	Reset the conversation for a new task
/compact [instructions]	Summarize and continue, keeping what you name
/context	See what is consuming the context window
/rewind	Return to an earlier point in conversation and code
/resume	Pick up a previous session
/model	Switch models
/permissions	Manage allow and deny rules
/hooks	Configure hooks
/agents	Create and manage subagents
/mcp	Manage MCP connections and authentication
/plugin	Browse, install, and manage plugins
/security-review	Audit pending changes for vulnerabilities
/review	Review the current diff, a branch, or a pull request (alias of /code-review)
/help	List commands and help

Twenty custom commands worth adding

Create each as a skill folder (.claude/skills/<name>/SKILL.md) so it works both as /name and automatically. Bodies below are the instructions; add frontmatter with a name and description. $ARGUMENTS receives whatever you type after the command.

Command	What it does
/standup	Summarize my commits and PR activity since yesterday as three bullets: done, next, blocked
/handoff	Write current status, decisions, and next steps to the plan file for the next session
/explain $ARGUMENTS	Explain how this file or flow works, with a call graph and the three riskiest spots
/spec $ARGUMENTS	Turn a rough idea into a spec with scope, constraints, edge cases, and definition of done
/plan	Explore, then write a step plan to docs/plans/; make no edits
/tdd $ARGUMENTS	Write failing tests for the spec, stop for review, then implement without editing tests
/fix-issue $ARGUMENTS	Read the ticket, reproduce in a test, fix, and reference the ticket in the commit
/grill	Ask the five hardest questions a skeptical staff engineer would ask about this branch
/break-it	Try to find inputs that break the changed functions; add tests for any found
/threat-model $ARGUMENTS	STRIDE-style threat model for a feature: assets, entry points, threats, mitigations
/audit-deps	Run the dependency-audit skill and summarize risk
/postmortem	Run the incident-postmortem skill on the linked incident
/release-notes $ARGUMENTS	Generate release notes between two refs
/api-docs	Update OpenAPI spec and examples from the current routes
/migration-plan $ARGUMENTS	Run the migration-plan skill for the named change
/flaky $ARGUMENTS	Classify a flaky test's likely cause and propose the minimal deterministic fix
/explain-plan	Explain the current Terraform plan, highlighting destroy and replace actions
/cleanup	Remove dead code, unused imports, and debug logs from the branch; no behavior changes
/refine	Knowing what was learned, propose the cleaner implementation of this branch
/onboard	Produce a one-page guide to this repository for a new engineer

Example: /fix-issue

---
name: fix-issue
description: Fixes a tracked issue end to end. Use when the user gives an issue or
  ticket ID, or says fix ENG-123, fix issue 42, or work on this ticket.
argument-hint: [issue-id]
---
Current branch: !`git branch --show-current`

1. Fetch issue $ARGUMENTS with the issue tracker tools. Summarize it in two lines.
2. Find the relevant code. State the suspected root cause with evidence.
3. Write a test that reproduces the bug. Confirm it fails.
4. Fix the code. Confirm the test passes and the suite is green.
5. Commit as `fix(scope): summary (refs $ARGUMENTS)`.
6. Draft a PR description using the pr-description skill.