Command reference

All dora commands — flags, exit codes, and common mistakes on one page.

Single page for every dora / doraval command. Prefer Audit or Quickstart if you want a tutorial.

Aliases: dora and doraval are the same binary. One-shot: npx @hacksmith/doraval ….

Contents

• Global flags
• scan · review · fix · new · memory
• reconcile · sessions · config
• bump · providers · update · completions

Global flags

These work on most commands:

Exit codes (typical)

scan

dora (alias: dora scan) is the default entrypoint. Read-only diagnosis of agent context in the current directory (or --cwd). No writes, no LLM, no API key.

dora
dora scan

dora --format json | jq '.summary'
dora --ci --cwd "$GITHUB_WORKSPACE"

Mistakes: scan only reports — use fix / review next. “What this repo uses” is dora; packaging layout is providers.

review

dora review grades skills, memory principles, and related artifacts. Tiers that can run do run; LLM and sessions skip when unavailable unless required.

dora review [PATH]

PATH is a skill directory or project root (default .).

Tiers: (1) structure · (2) heuristics · (3) LLM · (4) sessions. Details: Review tiers.

Memory principles apply as a rubric. Weight 7+ violations are errors. Memory.

dora review ./skills/deploy
dora review --quick .
dora review --deep ./skills/deploy
dora review --all --quick --ci
dora review CLAUDE.md

Mistakes: --deep without a judge → exit 2 (use --quick in CI). Skipped LLM tier ≠ full pass — configure a judge with config. Review does not rewrite files — use fix.

fix

dora fix applies mechanical fixes from scan/review. Diffs first; asks before write unless --yes.

dora fix [PATH]

dora fix
dora fix ./skills/deploy
dora fix --dry-run
dora fix --yes
dora fix --brief

Mistakes: run dora or review first. Judgment items need --brief + hand edit, then re-review.

new

dora new scaffolds a skill, rule, agent, or plugin. Prefer interactive once; flags for agents/CI.

dora new [TYPE] [NAME]

dora new
dora new skill --for claude --name review-pr --description "Reviews PRs" --yes
dora new rule --for cursor --name no-default-exports --yes
dora new plugin --for claude --name my-plugin --yes

No separate dora claude new path — always --for <agent>. Review the new path after scaffold.

memory

dora memory stores short, weighted principles. Review uses them as a rubric. Optional promote → AGENTS.md. Optional sync → private git backup. Old journal is gone (first run migrates once).

dora memory <subcommand>

memory add

dora memory add "Never use default exports" --weight 8
dora memory add "Run bun test before commit" --tags style,ci --body "CI fails otherwise"
dora memory add "Prefer named exports" --global

Title: short imperative, ≤ 80 characters.

memory list

dora memory list
dora memory list --format json

memory context

dora memory context
dora memory context --full
dora memory context --json --quiet
dora memory context --append-to CLAUDE.md

memory promote

dora memory promote
dora memory promote --weight 7 --dry-run
dora memory promote --yes

Default promote threshold weight 7+.

memory stash / restore

dora memory stash path/to/notes.md
dora memory restore path/to/notes.md
dora memory restore --dry-run

memory sync

dora memory sync
dora memory sync --repo owner/dora-memory

Mistakes: write enforceable principles, not essays. Soft weights rarely fail review. Update SessionStart hooks to dora memory context --json. Concept: Memory.

reconcile

Settles cross-agent contradictions (e.g. Claude vs Cursor) into shared AGENTS.md.

dora reconcile
dora reconcile --dry-run
dora reconcile --apply
dora reconcile --apply --yes

Workflow: dora → reconcile --dry-run → apply → re-scan/review.

Mistakes: always dry-run large monorepos first. Unused agent surfaces should be removed, not merged. Reconcile resolves conflicts — it does not invent policy.

sessions

Lists recent agent sessions for this project; show prints a timeline. Read-only.

dora sessions
dora sessions --agent claude --limit 20
dora sessions show <sessionId>

Adapters: Claude Code, Grok, Cursor, Codex, Copilot (when logs exist on disk).

dora review --sessions mines these logs for skill-usage evidence; requires recent sessions or exit 2.

Mistakes: empty list is normal on new machines. Not a live chat UI.

config

Reads and writes doraval settings (judge, model, agent command).

dora config                 # interactive hub (TTY)
dora config setup           # judge wizard
dora config get
dora config get eval.model
dora config set eval.model gpt-4.1
dora config set eval.judge auto

Common keys: eval.model, eval.provider, eval.judge, agent.command. Human get masks secrets; use --format json for raw values.

Mistakes: --deep needs a judge (config setup). Do not paste API keys into chat logs.

bump

Semver for plugin / marketplace manifests.

dora bump
dora bump patch
dora bump minor ./plugins
dora bump 1.2.3 --yes
dora bump --only plugin

Mistakes: pass the package root, not a random monorepo path. Review before ship.

providers

Static packaging/spec matrix (manifest paths, marketplace, MCP filename). Not “which agents this repo uses” — that is scan.

dora providers
dora providers --json

Scaffold with dora new --for <agent>.

update

Self-update doraval.

dora update
dora update --check
dora update --yes
dora update --via npm

See Installation.

completions

Shell tab completion (install plumbing, not a product subcommand). No dora completion command.

dora --completion bash
dora --completion zsh
dora --completion fish

Wire the printed script into your shell’s completion directory or rc file.

See also

• Audit my agent context
• Quickstart
• Use with your agent
• Review tiers
• Memory
• scenarios.yaml