Skip to content

Commit ed5a47a

Browse files
zrosenbauerclaude
zrosenbauer
and
claude
authored
feat(community): add roadmap docs and claude skills (#1)
Signed-off-by: Zac Rosenbauer <zacrosenbauer@gmail.com> Co-authored-by: Claude <noreply@anthropic.com>
1 parent 17534bf commit ed5a47a

36 files changed

Lines changed: 855 additions & 0 deletions

‎.claude/settings.json‎

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,5 @@
1+
{
2+
"enabledPlugins": {
3+
"plugin-dev@claude-plugins-official": true
4+
}
5+
}

‎.claude/skills/feature/SKILL.md‎

Lines changed: 104 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,104 @@
1+
---
2+
description: Add a new item to the product roadmap. Creates the feature markdown file, updates the overview table, searches for existing GitHub issues, and optionally creates a new issue.
3+
disable-model-invocation: true
4+
---
5+
6+
# Roadmap Item Management
7+
8+
Add new items to the product roadmap at `docs/roadmap/`. Each roadmap item consists of a feature markdown file and a corresponding row in the overview table, optionally linked to a GitHub issue.
9+
10+
## Workflow
11+
12+
### Step 1: Gather Information
13+
14+
Collect the following from the user before proceeding:
15+
16+
| Field | Required | Notes |
17+
|-------|----------|-------|
18+
| Title | Yes | Human-readable feature name |
19+
| Status | Yes | One of: `Released`, `Planned`, `Approved` |
20+
| Summary | Yes | One-sentence description of the feature |
21+
| Problem | Yes | What limitation or gap this addresses |
22+
| Solution | Yes | How the feature solves the problem |
23+
| Impact | Yes | What users gain from this feature |
24+
| GitHub Issue | No | Whether to search for / create a GitHub issue |
25+
26+
Use `AskUserQuestion` to collect any missing fields. All six content fields (Title, Summary, Problem, Solution, Impact, and Status) are required before creating the file.
27+
28+
### Step 2: Search for Existing Issues
29+
30+
Before creating a new issue, search for existing ones to avoid duplicates:
31+
32+
```bash
33+
gh issue list --repo joggrdocs/code --search "<title or keywords>" --state all --limit 10
34+
```
35+
36+
Present any matches to the user and ask whether to link an existing issue or create a new one.
37+
38+
### Step 3: Create GitHub Issue (if requested)
39+
40+
When the user wants a new issue, create it with the `enhancement` label:
41+
42+
```bash
43+
gh issue create \
44+
--repo joggrdocs/code \
45+
--title "<Title>" \
46+
--label "enhancement" \
47+
--body "$(cat <<'EOF'
48+
## Summary
49+
50+
<Summary text>
51+
52+
## Problem
53+
54+
<Problem text>
55+
56+
## Solution
57+
58+
<Solution text>
59+
60+
## Impact
61+
62+
<Impact text>
63+
EOF
64+
)"
65+
```
66+
67+
Capture the issue number from the output for use in the overview table.
68+
69+
### Step 4: Create Feature File
70+
71+
Generate the filename by converting the title to kebab-case (lowercase, hyphens for spaces, strip special characters). Write the file to `docs/roadmap/features/<kebab-case-name>.md`.
72+
73+
Use the format documented in `references/feature-template.md`. The status badge colors are:
74+
75+
| Status | Badge |
76+
|--------|-------|
77+
| Released | `![Released](https://img.shields.io/badge/Released-%23006400)` |
78+
| Planned | `![Planned](https://img.shields.io/badge/Planned-%23003366)` |
79+
| Approved | `![Approved](https://img.shields.io/badge/Approved-%23665500)` |
80+
81+
### Step 5: Update Overview Table
82+
83+
Append a new row to the table in `docs/roadmap/overview.md`. The row format is:
84+
85+
```
86+
| [<Title>](./features/<kebab-case-name>.md) | <Summary> | ![<Status>](https://img.shields.io/badge/<Status>-<color>) | <issue-link-or-dash> |
87+
```
88+
89+
Column order is: Feature, Description, Status, Issue.
90+
91+
- For the issue column, use `[#N](https://github.com/joggrdocs/code/issues/N)` if an issue exists, otherwise use `-`.
92+
- Insert the new row in the correct status group: Released items first, then Planned, then Approved.
93+
94+
### Step 6: Confirm
95+
96+
After creating the files, display a summary:
97+
98+
- Feature file path
99+
- Overview table updated
100+
- GitHub issue link (if created)
101+
102+
## Reference Files
103+
104+
- **`references/feature-template.md`** — Complete feature file template with correct badge format

‎.claude/skills/feature/references/feature-template.md‎

Lines changed: 94 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,94 @@
1+
# Feature File Template
2+
3+
Use this exact format when creating new roadmap feature files at `docs/roadmap/features/<kebab-case-name>.md`.
4+
5+
## Template
6+
7+
```markdown
8+
# <Title>
9+
10+
**Status:** ![<Status>](https://img.shields.io/badge/<Status>-<color>)
11+
12+
## Problem
13+
14+
<What limitation, gap, or pain point this feature addresses.>
15+
16+
## What we're releasing
17+
18+
<How the feature solves the problem — the approach or architecture.>
19+
20+
## Expected outcome
21+
22+
<What users gain — the concrete benefit or outcome.>
23+
```
24+
25+
## Status Badges
26+
27+
Use these exact badge URLs based on feature status:
28+
29+
### Released (dark green)
30+
31+
```
32+
![Released](https://img.shields.io/badge/Released-%23006400)
33+
```
34+
35+
### Planned (dark blue)
36+
37+
```
38+
![Planned](https://img.shields.io/badge/Planned-%23003366)
39+
```
40+
41+
### Approved (dark yellow/gold)
42+
43+
```
44+
![Approved](https://img.shields.io/badge/Approved-%23665500)
45+
```
46+
47+
## Example: Agent Harness
48+
49+
```markdown
50+
# Agent Harness
51+
52+
**Status:** ![Planned](https://img.shields.io/badge/Planned-%23003366)
53+
54+
## Problem
55+
56+
In the first release, the Coding Agent Setup Doctor used a single agent to spot-check repositories for AI development setup issues, which limited the depth and consistency of findings.
57+
58+
## What we're releasing
59+
60+
Agent Harness introduces a coordinated agentic loop—similar to architectures used by Claude Code, Codex, and other coding agent platforms—allowing multiple specialized sub-agents to analyze repositories more thoroughly and validate findings collaboratively.
61+
62+
## Expected outcome
63+
64+
Users receive significantly more comprehensive and accurate setup diagnostics, with a system designed to continuously improve over time and serve as the foundation for reliable AI coding workflows.
65+
```
66+
67+
## Filename Convention
68+
69+
Convert the feature title to kebab-case:
70+
71+
1. Lowercase all characters
72+
2. Replace spaces with hyphens
73+
3. Remove special characters (parentheses, colons, etc.)
74+
4. Collapse multiple hyphens into one
75+
76+
Examples:
77+
78+
| Title | Filename |
79+
|-------|----------|
80+
| Agent Harness | `agent-harness.md` |
81+
| Coding Agent Setup Doctor | `coding-agent-setup-doctor.md` |
82+
| GG Workflow | `gg-workflow.md` |
83+
| Coding Agent Toolkit MCP (Serena) | `coding-agent-toolkit-mcp.md` |
84+
| Secure Data Access Layer (unmcp) | `secure-data-access-layer.md` |
85+
86+
## Overview Table Row Format
87+
88+
```
89+
| [<Title>](./features/<filename>) | <Summary> | ![<Status>](https://img.shields.io/badge/<Status>-<color>) | <issue> |
90+
```
91+
92+
Where `<issue>` is either:
93+
- `[#N](https://github.com/joggrdocs/code/issues/N)` — linked GitHub issue
94+
- `-` — no issue yet

‎docs/roadmap/features/agent-harness.md‎

Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
# Agent Harness
2+
3+
**Status:** ![Planned](https://img.shields.io/badge/Planned-%23003366)
4+
5+
## Summary
6+
7+
Sub-agent orchestration setup for Joggr Agents using the same agentic loop architecture as Claude Code, Codex, OpenCode and other coding agent platforms.
8+
9+
## Problem
10+
11+
In the first release, the Coding Agent Setup Doctor used a single agent to spot-check repositories for AI development setup issues, which limited the depth and consistency of findings.
12+
13+
## Solution
14+
15+
Agent Harness introduces a coordinated agentic loop—similar to architectures used by Claude Code, Codex, and other coding agent platforms—allowing multiple specialized sub-agents to analyze repositories more thoroughly and validate findings collaboratively.
16+
17+
## Impact
18+
19+
Users receive significantly more comprehensive and accurate setup diagnostics, with a system designed to continuously improve over time and serve as the foundation for reliable AI coding workflows.

‎docs/roadmap/features/agents-skills-registry-global.md‎

Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
# Agents & Skills Registry: Global
2+
3+
**Status:** ![Approved](https://img.shields.io/badge/Approved-%23006B3F)
4+
5+
## Summary
6+
7+
A set of prebuilt and Joggr-verified agents and reusable skills installable via the AI Dev Setup CLI, scoped to repository needs.
8+
9+
## Problem
10+
11+
Developers configuring AI coding agents must manually discover, vet, and wire up individual agent capabilities, leading to inconsistent setups and duplicated effort across repositories.
12+
13+
## Solution
14+
15+
The Global Registry provides a curated catalog of prebuilt, Joggr-verified agents and reusable skills that can be installed directly through the AI Dev Setup CLI and automatically scoped to a repository's needs.
16+
17+
## Impact
18+
19+
Teams can bootstrap reliable AI coding workflows in minutes instead of hours, with confidence that every agent and skill has been verified for quality and compatibility.

‎docs/roadmap/features/agents-skills-registry-private.md‎

Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
# Agents & Skills Registry: Private
2+
3+
**Status:** ![Approved](https://img.shields.io/badge/Approved-%23006B3F)
4+
5+
## Summary
6+
7+
A private hosted registry of skills searchable and accessible by humans and coding agents, custom to the enterprise, used to limit maintenance, enforce security, and understand distribution and usage across the enterprise.
8+
9+
## Problem
10+
11+
Enterprises lack a centralized, private way to manage custom agent skills, leading to skill sprawl, security blind spots, and no visibility into which skills are used where across the organization.
12+
13+
## Solution
14+
15+
The Private Registry gives enterprises a hosted, searchable catalog of custom skills that enforces security policies, reduces maintenance overhead, and provides full visibility into skill distribution and usage.
16+
17+
## Impact
18+
19+
Enterprise teams gain centralized governance over their AI coding toolchain, reducing security risk while enabling consistent, organization-wide adoption of approved agent skills.

‎docs/roadmap/features/coding-agent-quality-checks.md‎

Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
# Coding Agent Quality Checks
2+
3+
**Status:** ![Approved](https://img.shields.io/badge/Approved-%23006B3F)
4+
5+
## Summary
6+
7+
Automatically installs and maintains Coding Agent (Claude Code) hooks that run linting, formatting, and CLI-based validation on any code an agent touches, ensuring agent-generated code meets the same quality gates as human-written code.
8+
9+
## Problem
10+
11+
AI coding agents can generate code that bypasses a project's established linting, formatting, and validation rules, introducing inconsistencies and quality regressions that are costly to catch in review.
12+
13+
## Solution
14+
15+
Coding Agent Quality Checks automatically installs and maintains hooks that run the project's full suite of linting, formatting, and CLI-based validation on every piece of agent-generated code before it is committed.
16+
17+
## Impact
18+
19+
Teams can trust that agent-generated code consistently meets the same quality standards as human-written code, reducing review burden and preventing quality regressions from entering the codebase.

‎docs/roadmap/features/coding-agent-setup-cli.md‎

Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
# Coding Agent Setup CLI
2+
3+
**Status:** ![Planned](https://img.shields.io/badge/Planned-%23003366)
4+
5+
## Summary
6+
7+
Guides users through reviewing their current and establishing their (harden existing) AI development setup, generating initial versions of files such as `CLAUDE.md`, `AGENTS.md`, and `.claude/settings.json` based on detected tools and repository context.
8+
9+
## Problem
10+
11+
Setting up AI coding agent configurations from scratch requires deep knowledge of each agent platform's file formats and best practices, creating a high barrier to entry for teams adopting AI-assisted development.
12+
13+
## Solution
14+
15+
The Setup CLI provides an interactive, guided workflow that detects existing tools and repository context to automatically generate properly configured instruction files and settings for AI coding agents.
16+
17+
## Impact
18+
19+
Teams can bootstrap a production-ready AI development setup in minutes, reducing onboarding friction and ensuring consistent, best-practice configurations across repositories.

‎docs/roadmap/features/coding-agent-setup-doctor.md‎

Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
# Coding Agent Setup Doctor
2+
3+
**Status:** ![Released](https://img.shields.io/badge/Released-%23008A00)
4+
5+
## Summary
6+
7+
Analyzes a repository for misconfigured AI development setup, including poorly written/outdated instruction files and incorrect/unsafe tool configurations required by AI coding agents.
8+
9+
## Problem
10+
11+
Repositories often have misconfigured, outdated, or insecure AI agent settings that silently degrade coding agent performance and can introduce unsafe tool permissions.
12+
13+
## Solution
14+
15+
The Setup Doctor scans a repository's AI development configuration end-to-end, identifying poorly written instruction files, outdated settings, and incorrect or unsafe tool configurations in a single diagnostic pass.
16+
17+
## Impact
18+
19+
Users gain immediate visibility into the health of their AI development setup, enabling them to fix issues before they affect coding agent reliability or security.

‎docs/roadmap/features/coding-agent-setup-pr-check.md‎

Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
# Coding Agent Setup PR Check
2+
3+
**Status:** ![Planned](https://img.shields.io/badge/Planned-%23003366)
4+
5+
## Summary
6+
7+
Detect and flag outdated/misconfigured agent configs and instruction files on every pull request.
8+
9+
## Problem
10+
11+
Agent configuration and instruction files can silently become outdated or broken as a codebase evolves, and these issues often go unnoticed until they cause incorrect AI behavior in production workflows.
12+
13+
## Solution
14+
15+
The PR Check runs automatically on every pull request to validate agent configs and instruction files, flagging any that are outdated or misconfigured before they are merged.
16+
17+
## Impact
18+
19+
Teams catch AI development setup regressions early in the development cycle, preventing broken agent configurations from reaching the main branch and disrupting AI-assisted workflows.

0 commit comments

Comments
 (0)