Quickstart
End-to-end tutorial — install the skill, install the CLI, scan, review, fix, scaffold, and remember principles.
Outcome: doraval skill installed, dora available, project scanned and reviewed, and you know the main commands.
Short path only (skill + review)? Use Audit my agent context instead.
What you will do
- Install the doraval skill
- Install the CLI (or keep using
npx) - Scan the project
- Review and fix
- Scaffold new context (optional)
- Record a principle (optional)
Time: about 10–15 minutes. Needs: Node (for npx), a project with or without agent files.
1. Install the doraval skill
npx skills add saif-shines/doraval
Agents load this checklist when you edit skills, plugins, rules, or agent config. The skill runs the same dora engine you use by hand (or npx @hacksmith/doraval if dora is not on PATH).
What if npx skills add fails?
- Need network and a current Node
- Retry:
npx skills add saif-shines/doraval - List installed skills:
npx skills list
2. Put the CLI on PATH (or use npx)
One-shot (no install):
npx @hacksmith/doraval --version
Permanent (pick one):
# macOS
brew tap saif-shines/tap && brew trust saif-shines/tap && brew install doraval
# npm
npm install -g @hacksmith/doraval
# Bun
bun add -g @hacksmith/doraval
After install, dora and doraval are the same binary. Verify:
dora --version
PATH / Alpine notes
- Global install but
doranot found: add your npm/Bun global bin directory toPATH - Alpine/musl: use Bun (
bunx/bun add -g); platform binaries target glibc
Full matrix: Installation.
3. Scan the project
dora
# same as: dora scan
# or: npx @hacksmith/doraval
Read-only. No API key. You get:
| Section | Meaning |
|---|---|
| Agent surfaces | Which agents look configured |
| Health | Skill / manifest issues |
| Next | Numbered follow-ups (dora fix …, dora review …) |
dora --format json # for agents / scripts
dora --ci # implies JSON
4. Review quality
dora review .
dora review ./path/to/skill-dir
dora review --all
dora review --quick . # structure + heuristics only
dora review --all --quick --ci # CI gate, no LLM
| Exit | Meaning |
|---|---|
0 |
Clean |
1 |
Issues found |
2 |
Could not run |
Tiers (structure → heuristics → LLM → sessions): Review tiers.
5. Fix mechanical issues
dora fix .
dora fix --dry-run
dora fix --yes # agents / CI (no TTY prompt)
dora fix --brief # prompt for judgment items
| Kind | Examples | Behavior |
|---|---|---|
| Mechanical | Frontmatter, paths | Diff first; apply on approval or --yes |
| Judgment | Vague descriptions | Not auto-written; use --brief then re-review |
Re-run dora review . until exit 0.
6. Scaffold new context (optional)
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
| Type | Meaning |
|---|---|
skill |
Reusable SKILL.md |
rule |
Always-on convention |
agent |
Subagent role |
plugin |
Package to ship |
Always pass --for <agent> for non-interactive runs. Then dora review the new path.
7. Record a principle (optional)
dora memory add "Never use default exports" --weight 8
dora memory list
dora memory promote # hard rules → AGENTS.md (diff + confirm)
dora memory context --json --quiet # SessionStart hooks
Weights: 8+ hard rule · 5 default · 1–3 soft. Memory concept.
8. Cross-agent conflicts (when scan reports them)
dora reconcile --dry-run
dora reconcile
# or: dora reconcile --apply --yes
Settles contradictions into shared AGENTS.md. dora reconcile.
Use with your agent
After step 1, paste or rely on the skill:
Review my agent context with dora (npx @hacksmith/doraval). Fix mechanical issues with --yes. Do not report done until review exits 0.
JSON, exit codes, CI examples: Use with your agent.
Command map
| Job | Command |
|---|---|
| Scan project | dora |
| Review | dora review |
| Fix | dora fix |
| Scaffold | dora new --for … |
| Principles | dora memory |
| Conflicts | dora reconcile |
| Sessions | dora sessions |
| Config / judge | dora config |
Full flags: Command reference.
Next
- Audit my agent context — short skill + review path only
- Use with your agent — automation details
- Command reference — full flags for every command