tessl i github:cteyton/packmind --skill packmind-onboardComplete automated onboarding: analyzes codebase, creates package, and generates standards & commands via CLI. Automatic package creation when none exist, user selection when packages are available.
Action skill. Provides complete automated onboarding for Packmind:
Automatic package creation when none exist, user selection when packages are available.
-2, -3, etc.path[:line-line] entries; omit line ranges only when the file isn't text-searchable.Print exactly:
I'll start the Packmind onboarding process. I'll create your first standards and commands and send them to your Packmind organization. This usually takes ~3 minutes.Get the repository name for package naming:
basename "$(git rev-parse --show-toplevel)"Remember this as the repository name for package creation in Step 2.
Also run packmind-cli whoami and extract the Host: value from the output. Remember this URL for the completion summary.
Handle package creation or selection.
List available packages:
packmind-cli install --listParse the output to get package names.
Auto-create package using repository name:
packmind-cli packages create "${REPO_NAME}-standards" --origin-skill=onboardPrint:
No existing packages found — created a new one: ${REPO_NAME}-standardsAsk via AskUserQuestion:
{package-name}?"Ask via AskUserQuestion:
${REPO_NAME}-standards as default)packmind-cli packages create <name> --origin-skill=onboardRemember the selected/created package name for later reference.
Print exactly:
packmind-onboard: analyzing codebase (read-only)
Target package: [package-name]Before analyzing, detect and preserve any existing Packmind/agent configuration.
Glob for markdown in these roots (recursive):
.packmind/**/*.md.claude/**/*.md.agents/**/*.md**/skills/**/*.md**/rules/**/*.mdClassify found files into counts:
.packmind/standards/**/*.md.packmind/commands/**/*.md.claude/, .agents/, or any skills/ or rules/ directory outside .packmindIf any exist, print exactly:
Existing Packmind/agent docs detected:
Standards: [N]
Commands: [M]
Other docs: [P]No overwrites. New files (if you Export) will be added next to the existing ones.
package.json, pnpm-lock.yaml, yarn.lock, tsconfig.jsonpyproject.toml, requirements.txt, setup.pygo.modCargo.tomlGemfilepom.xml, build.gradle, build.gradle.kts*.csproj, *.slncomposer.jsonsrc/application/, src/domain/, src/infra/src/controllers/, src/services/packages/, apps/Print exactly:
Stack detected (heuristic):
Languages: [..]
Repo shape: [monorepo|single]
Architecture markers: [..|none]Read each reference file for detailed search patterns, thresholds, and insight templates.
| Analysis | Reference File | Output focus |
|---|---|---|
| File Template Consistency | references/file-template-consistency.md | Commands |
| CI/Local Workflow Parity | references/ci-local-workflow-parity.md | Commands |
| Role Taxonomy Drift | references/role-taxonomy-drift.md | Standards |
| Test Data Construction | references/test-data-construction.md | Standards |
For every finding, keep an internal record:
INSIGHT:
title: ...
why_it_matters: ...
confidence: [high|medium|low]
evidence:
- path[:line-line]
where_it_doesnt_apply:
- path[:line-line]Generate all draft files in one batch, using the formats defined above.
For each Standard insight, create a Markdown file at .packmind/standards/_drafts/<slug>.draft.md:
# Standard Name
What the standard covers and why.
## Scope
Where this standard applies (e.g., 'TypeScript files', 'React components').
## Rules
### Rule starting with action verb
Another rule can follow...
## Examples
### Good
```typescript
// Valid code example
```// Invalid code example### Command Draft Format
For each Command insight, create a Markdown file at `.packmind/commands/_drafts/<slug>.draft.md`:
```markdown
# Command Name
What the command does, why it's useful, and when it's relevant.
## When to Use
- Scenario when this command applies
- Another scenario...
## Checkpoints
- Question to validate before proceeding?
## Steps
### 1. Step Name
What this step does and how to implement it.
```typescript
// Optional code exampleDescription of next step...
### Generation Rules
- Generate drafts **only from discovered insights** (no invention)
- Use evidence from analysis to populate rules/steps
- Cap output: max **5 Standards** + **5 Commands**
- Never overwrite existing files; append `-2`, `-3`, etc. if slug exists
---
## Step 8 — Present Summary & Confirm
Present the generated draft files and ask for confirmation:Target package: [package-name] Stack detected: [languages], [monorepo?], [architecture markers] Analyses run: [N] checks
DRAFTS CREATED:
Standards ([N]):
Commands ([M]):
Then ask via AskUserQuestion with three options:
- **Create all now** — Proceed with creating all standards and commands
- **Let me review drafts first** — Pause to allow editing, re-run skill when ready
- **Cancel** — Exit without creating anything
---
## Step 9 — Create Items
### If user selected "Create all now"
**IMPORTANT:** The CLI only accepts JSON playbook files, not markdown. Before calling the CLI, convert each `.draft.md` file to a `.json` file.
#### Standard JSON Schema
Convert the markdown draft to this JSON format:
```json
{
"name": "Standard name (from # heading)",
"description": "What the standard covers (from intro paragraph)",
"scope": "Where it applies (from ## Scope section)",
"rules": [
{
"content": "Rule starting with action verb (from ### Rule headings under ## Rules)",
"examples": {
"positive": "Valid code example (from ### Good section)",
"negative": "Invalid code example (from ### Bad section)",
"language": "TYPESCRIPT"
}
}
]
}Convert the markdown draft to this JSON format:
{
"name": "Command name (from # heading)",
"summary": "What it does and when (from intro paragraph)",
"whenToUse": ["Scenario 1", "Scenario 2 (from ## When to Use bullets)"],
"contextValidationCheckpoints": ["Question 1? (from ## Checkpoints bullets)"],
"steps": [
{
"name": "Step name (from ### N. Step Name)",
"description": "Step description (from step content)",
"codeSnippet": "Optional code fence content"
}
]
}For each standard draft:
.draft.md file.packmind/standards/_drafts/<slug>.jsonpackmind-cli standards create .packmind/standards/_drafts/<slug>.json --origin-skill=onboardpackmind-cli packages add --to <package-slug> --standard <slug> --origin-skill=onboardFor each command draft:
.draft.md file.packmind/commands/_drafts/<slug>.jsonpackmind-cli commands create .packmind/commands/_drafts/<slug>.json --origin-skill=onboardpackmind-cli packages add --to <package-slug> --command <slug> --origin-skill=onboardShow progress:
Sending standards and commands to your Packmind organization...
✓ error-handling-pattern
✓ naming-conventions
✗ test-factory-patterns (error: duplicate name exists)
✓ run-full-test-suite
Done: 3 created, 1 failedPrint:
Draft files ready for review at:
- .packmind/standards/_drafts/
- .packmind/commands/_drafts/
Edit them as needed, then re-run this skill to continue.Exit the skill.
Print:
Onboarding cancelled.
Draft files remain at .packmind/*/_drafts/ if you want to review them later.Exit the skill.
============================================================
✅ ONBOARDING COMPLETE
============================================================
Package: [package-name]
Created: [N] standards, [M] commands
Your standards and commands have been created and deployed locally.
Next steps:
- Reload your AI coding assistant to start using them
- Visit [host from packmind-cli whoami] to manage your standards and commands
- Run `packmind-cli install [package-slug]` in other repos to distribute them
============================================================Clean up successful draft files after creation.
============================================================
⚠️ ONBOARDING COMPLETED WITH ERRORS
============================================================
Package: [package-name]
Created: [N] standards, [M] commands
Failed: [X] items
Failed items:
• [item-name]: [error message]
Failed drafts remain in .packmind/*/_drafts/ for review.
You can fix and re-run, or create manually with:
packmind-cli standards create <file> --origin-skill=onboard
packmind-cli packages add --to <package-slug> --standard <slug> --origin-skill=onboard
packmind-cli commands create <file> --origin-skill=onboard
packmind-cli packages add --to <package-slug> --command <slug> --origin-skill=onboard
============================================================Keep failed draft files for user to fix and retry.
If analysis found no patterns:
============================================================
ℹ️ NO PATTERNS DISCOVERED
============================================================
The analysis didn't find enough recurring patterns to generate standards or commands.
This can happen with smaller codebases or projects with very diverse coding styles.
You can try again later as the codebase grows, or create standards manually with:
packmind-cli standards create <file> --origin-skill=onboard
============================================================If packmind-cli packages create fails:
❌ Failed to create package: [error message]
Please check:
- You are logged in: `packmind-cli login`
- Your network connection is working
- The package name is valid
Cannot proceed with onboarding until package is created.Exit the skill. Do not proceed to analysis.
If CLI commands fail with authentication errors:
❌ Not logged in to Packmind
Please run:
packmind-cli login
Then re-run this skill.If packmind-cli install --list returns no packages:
Auto-create a package using the repository name.
Since the onboard skill is present, the user has configured an AI agent. Deploy the created artifacts locally using the package selected/created in Step 2:
packmind-cli install <package-slug>This deploys to agent-specific folders:
| Agent | Standards | Commands |
|---|---|---|
| Claude | .claude/rules/packmind/standard-[slug].md | .claude/commands/packmind/[slug].md |
| Cursor | .cursor/rules/packmind/standard-[slug].mdc | .cursor/commands/packmind/[slug].mdc |
| Copilot | .github/instructions/packmind-standard-[slug].instructions.md | .github/prompts/packmind-[slug].prompt.md |
Delete the draft files, then print final summary:
============================================================
PUBLISHED & DEPLOYED
============================================================
Standards and commands have been sent to your Packmind organization
and deployed to your AI coding assistant's configuration files.
Standards: [N]
- [Name] (slug: [slug])
→ .packmind/standards/[slug].md
→ [agent-specific path]
Commands: [M]
- [Name] (slug: [slug])
→ .packmind/commands/[slug].md
→ [agent-specific path]
Draft files cleaned up.
============================================================If user declines (N):
Print:
Draft files ready for review at:
- .packmind/standards/_drafts/
- .packmind/commands/_drafts/
Edit them as needed, then re-run this skill to create them.If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.